Public DARPA programs on

darpa.mil

Run through GPT-3 for simple summaries

Logan Graham, October 2022

1
Reimagining Protein Manufacturing
(RPM)
Dr. Amy Jenkins

https://www.darpa.mil/program/reimagining-protein-manufacturing.html

GPT-3 Summary: The goal of the RPM program is to develop a faster and more ef-
ficient way to produce protein-based medical countermeasures (MCMs). The program
will address three challenges: (1) achieving relevant quantities of MCM; (2) ensuring
initiation of protein production within 24 hours of receiving DNA/RNA template;
and (3) producing fully bioactive complex proteins by ensuring correct folding and
controlled addition of post-translational modifications. The RPM program seeks to
revolutionize the production of protein-based MCMs and lay the foundation for the
fully distributed manufacturing paradigms of the future.

Full Details: DoD access to critical proteins is currently limited by the slow response
times of protein production methods. State-of-the-art (SOA) protein-based medical
countermeasure (MCM) production relies on massive centralized infrastructure and
complex pipelines that require lengthy cellular engineering, sufficient growth of cells
for production, and intensive purification and quality control. Current response times
are several months for production alone, with additional time needed for dispersing
MCMs to vulnerable populations. Distributed manufacturing paradigms may trans-
form these timelines into those relevant to DoD response, especially in denied, de-
graded, or disrupted operational environments. To significantly advance the DoD’s
ability to rapidly manufacture proteins in a fully distributed setting, RPM will ad-
dress the following biochemistry challenges: (1) achieving relevant quantities of MCM;
(2) ensuring initiation of protein production within ˜24 hours of receiving DNA/RNA
template; and (3) producing fully bioactive complex proteins by ensuring correct fold-
ing and controlled addition of post-translational modifications. The RPM program
seeks to revolutionize the production of protein based MCMs and lay the foundation
for the fully distributed manufacturing paradigms of the future.

2
Hardening Development Toolchains
Against Emergent Execution Engines
(HARDEN)
Dr. Sergey Bratus

https://www.darpa.mil/program/hardening-development-toolchains-against-emergent-
execution-engines.html

GPT-3 Summary: GOAL: The goal of this program is to develop theories and
approaches to anticipate, isolate, and mitigate emergent behaviors in computing sys-
tems throughout the entire software development lifecycle.
METHODS: This program will explore novel theories and approaches, and seek to
develop practical tools to anticipate, isolate, and mitigate emergent behaviors in com-
puting systems throughout the entire software development lifecycle.
IMPACT: The goal is to integrate those capabilities into the standard processes of
the software development lifecycle.

Full Details: The Department of Defense (DoD) has a critical need to deny cyber
attackers the capability to execute unintended, yet robust and often unobservable
computations on DoD systems and critical infrastructure systems. Empirically, mod-
ern exploitation methods rely on long chains of emergent behaviors of the target’s
unprotected computational abstractions, where attackers leverage one combination
of abstractions to create an ephemeral state in which the next set of unprotected ab-
stractions is exposed, until the goals of exploitation are achieved. Counterintuitively,
instead of being brittle and easily disrupted, these chains are robust and portable be-
tween implementations independently created by different vendors. This phenomenon
is colloquially described as “weird machines”—well-defined, robust, and abstract-able
engines of emergent execution (EE) and adversarial programmability. These machines
are often unintentionally programmed into the target and are merely unlocked for an
attacker’s use through coding flaws.The Hardening Development Toolchains Against
Emergent Execution Engines (HARDEN) program will explore novel theories and
approaches, and seeks to develop practical tools to anticipate, isolate, and mitigate
emergent behaviors in computing systems throughout the entire software develop-
ment lifecycle. HARDEN aims to radically improve security outcomes in software
for integrated systems by creating novel tools, metadata, and instrumentation for
emergent computation. It also seeks to efficiently mitigate exploitation of software
abstractions and protect intended abstractions from adversarial reuse. The goal is to

3
integrate those capabilities into the standard processes of the software development
lifecycle.Additional information is available in the HARDEN Program Announce-
ment.

4
Searchlight
Mr. John-Francis Mergen

https://www.darpa.mil/program/searchlight.html

GPT-3 Summary: GOAL: To develop techniques and systems that enable an en-
terprise to temporarily decrease the QoS for low-priority application traffic internal
to the organization, resulting in sufficient QoS for the organization’s high-priority
traffic.
METHODS:
IMPACT:

Full Details: Distributed applications are important tools for managing global en-
terprises as they improve both the speed and scale of decision-making, learning, and
other critical functions. Virtual documents are one example of a commonly used dis-
tributed application. These provide organizations with the ability to have multiple
writers and editors collaborate on document authoring in near real-time regardless
of their physical locations. These critical productivity tools rely on internet-enabled,
enterprise-wide communication systems to interconnect sites and create a global sub-
strate to support their operation. While these applications have become increasingly
vital for enterprises, approaches to specifying and controlling network quality of ser-
vice (QoS) for them has not been deployed at scale. There are many reasons for this,
including the complexity of decentralized resource control in a federated internet, and
the demonstrated effectiveness of low-level Internet protocols such as TCP/IP (Trans-
mission Control Protocol / Internet Protocol) in exploiting and sharing the available
capacity amongst networked applications. The Searchlight program explores novel
approaches to analysis and QoS management of an enterprise’s distributed applica-
tions overlaid on the internet. The goal of the program is to develop techniques and
systems that enable an enterprise to temporarily decrease the QoS for low-priority
application traffic internal to the organization, resulting in sufficient QoS for the or-
ganization’s high-priority traffic. More information is available in the Searchlight
BAA.

5
Foundations Required for Novel Com-
pute (FRANC)
Dr. Yogendra Joshi

https://www.darpa.mil/program/foundations-required-for-novel-compute.html

GPT-3 Summary: GOAL: To develop innovative approaches to advance compute

technologies beyond the Von Neumann topology.
TECHNOLOGY METHODS: Leveraging recent advances in materials, devices, and
integration technology, the program seeks to develop novel memory-centric compute
topologies that break the traditional separation of processors and memory components
to realize dramatic advances in compute efficiency and throughput of the workload,
especially for applications constrained by size, weight, and power (SWaP).
IMPACT: Such approaches are particularly suited for applications relevant to artifi-
cial intelligence (AI) where in-memory computation provides unique advantages over
traditional Von Ne

Full Details: The Von Neumann architecture has significantly aided the rapid ad-
vancement of computing over the past seven decades. However, moving data be-
tween the processors and memory components of this architecture requires significant
time and high-energy consumption, which constrains the computing performance and
workload. Overcoming this bottleneck requires new computing architectures and de-
vices that can significantly advance the computing performance beyond the traditional
practice of transistor scaling (i.e., Moore’s Law). The Foundations Required for Novel
Compute (FRANC) program aims to develop innovative approaches to advance com-
pute technologies beyond the Von Neumann topology. Leveraging recent advances
in materials, devices, and integration technology, the program seeks to develop novel
memory-centric compute topologies that break the traditional separation of proces-
sors and memory components to realize dramatic advances in compute efficiency and
throughput of the workload, especially for applications constrained by size, weight,
and power (SWaP). Innovative compute architectures and new, fast non-volatile stor-
age and memory-centric computing devices will be explored under FRANC to enable
low latency compute near or inside the data storage elements. Such approaches
are particularly suited for applications relevant to artificial intelligence (AI) where
in-memory computation provides unique advantages over traditional Von Neumann
computation.

6
Verified Security and Performance
Enhancement of Large Legacy Soft-
ware (V-SPELLS)
Dr. Sergey Bratus

https://www.darpa.mil/program/verified-security-and-performance-enhancement-of-large-
legacy-software.html

GPT-3 Summary: GOAL: The V-SPELLS program will create practical tools for
developers to gain benefits of formal software verification in incremental software
(re)engineering.

TECHNOLOGY METHODS:

The V-SPELLS program will leverage a combination of novel concepts in pro-

gram understanding and verification, as well as recent developments in domain
specific languages and composable system architectures for production systems at
scale.

IMPACT:

The V-SPELLS program seeks to radically broaden adoption of software veri-

fication by enabling incremental introduction of superior technologies into systems
that cannot be re-designed from scratch and

Full Details: The DoD has a critical need for enhancing and replacing components
of existing software with more secure and more performant code. This includes cases
where a key performance or security benefit comes from moving parts of the software
to new hardware, such as utilizing hardware accelerators, isolation enclaves, offload
processors, and distributed computation. However, introducing enhancements or re-
placements into large legacy code bases carries a high risk that new code will not safely
compose with the rest of the system. Verified programming methodologies for creat-
ing software that is correct-by-construction are currently not effective for lowering this
risk, because they focus on clean-slate software construction, assume an existing for-
mal specification that is typically not available for a legacy system, and require formal
methods expertise typically not accessible to developers.The goal of the V-SPELLS
program is to create a developer-accessible capability for piece-by-piece enhancement

7
of software components with new verified code that is both correct-by-construction
and compatible-by-construction, i.e., safely composable with the rest of the system.
V-SPELLS will create practical tools for developers to gain benefits of formal software
verification in incremental software (re)engineering rather than only in clean-slate in-
troduction. V-SPELLS tools will enable developers to deliver assured incremental
modernization of legacy systems in a manner that leverages verification technologies
and reduces rather than raises risk. V-SPELLS aims to radically broaden adoption
of software verification by enabling incremental introduction of superior technologies
into systems that cannot be re-designed from scratch and replaced as a whole.To
accomplish these goals, the V-SPELLS program will leverage a combination of novel
concepts in program understanding and verification, as well as recent developments
in domain specific languages and composable system architectures for production sys-
tems at scale. The program seeks breakthroughs in and novel approaches to the fol-
lowing technical challenges, including, but not limited to:Automated program under-
standing to infer architectural structure, assumptions, and dependencies in a legacy
source code base, enabling its decomposition into components with explicit modular
structure, interfaces, dependencies, and constraints. Recovery of domain abstrac-
tions and models from legacy code bases, in succinct and expressive representations
suitable for programming functional component enhancements or replacements that
are safely composable with the existing systems.Matching known and extracted do-
main abstractions and models with legacy code, lifting of legacy code to succinct,
enhanceable, safely composable, and inter-operating representations, and automated
specification inference leveraging such representation. Provably safe composition of
enhancements with the rest of the system.Overcoming performance reduction due
to added layers of abstraction with novel verified cross-layer optimization and dis-
tribution techniques (“verified stack flattening and distribution”). Development of
non-brittle and granular rules for composable representation, packaging, and trans-
formation of software verification proofs that support distribution and orchestration
of verified programs.The V-SPELLS BAA is available here.

8
Influence Campaign Awareness and
Sensemaking (INCAS)
Dr. Brian Kettler

https://www.darpa.mil/program/influence-campaign-awareness-and-sensemaking.html

GPT-3 Summary: The goal of the INCAS program is to develop techniques and
tools that will enable analysts to detect, characterize, and track geopolitical influence
campaigns with quantified confidence. The methods used will be based on social
listening and digital marketing, and the impact will be a more effective way of un-
derstanding and responding to these campaigns.

Full Details: The U.S. is engaged with its adversaries in an asymmetric, contin-
ual, war of weaponized influence narratives. Adversaries exploit misinformation and
true information delivered via influence messaging: blogs, tweets, and other online
multimedia content. Analysts require effective tools for continual sensemaking of the
vast, noisy, adaptive information environment to identify adversary influence cam-
paigns.Today, geopolitical influence campaign detection and sensemaking is largely
manual and ad hoc. Analysts use social listening tools to formulate complex keyword
queries; track trending keywords, hashtags, and topics; and read hundreds to thou-
sands of documents to identify influence themes. New or “low and slow” campaigns
are hard to detect early as their message volume may be beneath platform “trend-
ing” thresholds and pertinent hashtags may be unknown. With current tools, it is
difficult to connect messages over time and across multiple platforms to track evolv-
ing campaigns and to assess confidence in analytic conclusions in a rigorous manner.
Confidence assessment by analysts is ad hoc, manual, subjective, qualitative, and
susceptible to analyst cognitive biases (e.g., confirmation bias). Analyst reports of-
ten cover static time ranges, and static reports quickly become stale.With current
tools, analysts must manually sift through a high volume of messages to find those
with relevant influence agendas then gauge which ones are gaining traction and with
whom. Analysts track population response using digital marketing tools for analyzing
audience demographics, interests, and personality. These tools lack explanatory and
predictive power for deeper issues of geopolitical influence. Audience analysis is often
done using static, demographic segmentation based on online and survey data. This
lacks the flexibility, resolution, and timeliness needed for dynamic geopolitical influ-
ence campaign detection and sensemaking.The Influence Campaign Awareness and
Sensemaking (INCAS) program will develop techniques and tools that enable analysts
to detect, characterize, and track geopolitical influence campaigns with quantified

9
confidence.For more information, see the INCAS Broad Agency Announcement.

10
Influence Campaign Awareness and
Sensemaking (INCAS)
Dr. Brian Kettler

https://www.darpa.mil/program/influence-campaign-awareness-and-sensemaking.html

GPT-3 Summary: The goal of the INCAS program is to develop techniques and
tools that will enable analysts to detect, characterize, and track geopolitical influence
campaigns with quantified confidence. The methods used will be based on social
listening and digital marketing, and the impact will be a more effective way of un-
derstanding and responding to these campaigns.

Full Details: The U.S. is engaged with its adversaries in an asymmetric, contin-
ual, war of weaponized influence narratives. Adversaries exploit misinformation and
true information delivered via influence messaging: blogs, tweets, and other online
multimedia content. Analysts require effective tools for continual sensemaking of the
vast, noisy, adaptive information environment to identify adversary influence cam-
paigns.Today, geopolitical influence campaign detection and sensemaking is largely
manual and ad hoc. Analysts use social listening tools to formulate complex keyword
queries; track trending keywords, hashtags, and topics; and read hundreds to thou-
sands of documents to identify influence themes. New or “low and slow” campaigns
are hard to detect early as their message volume may be beneath platform “trend-
ing” thresholds and pertinent hashtags may be unknown. With current tools, it is
difficult to connect messages over time and across multiple platforms to track evolv-
ing campaigns and to assess confidence in analytic conclusions in a rigorous manner.
Confidence assessment by analysts is ad hoc, manual, subjective, qualitative, and
susceptible to analyst cognitive biases (e.g., confirmation bias). Analyst reports of-
ten cover static time ranges, and static reports quickly become stale.With current
tools, analysts must manually sift through a high volume of messages to find those
with relevant influence agendas then gauge which ones are gaining traction and with
whom. Analysts track population response using digital marketing tools for analyzing
audience demographics, interests, and personality. These tools lack explanatory and
predictive power for deeper issues of geopolitical influence. Audience analysis is often
done using static, demographic segmentation based on online and survey data. This
lacks the flexibility, resolution, and timeliness needed for dynamic geopolitical influ-
ence campaign detection and sensemaking.The Influence Campaign Awareness and
Sensemaking (INCAS) program will develop techniques and tools that enable analysts
to detect, characterize, and track geopolitical influence campaigns with quantified

11
confidence.For more information, see the INCAS Broad Agency Announcement.

12
Air Combat Evolution (ACE)
Lt. Col. Ryan Hefron

https://www.darpa.mil/program/air-combat-evolution.html

GPT-3 Summary: GOAL: To increase trust in combat autonomy by using human-

machine collaborative dogfighting as its challenge problem.
TECHNOLOGY METHODS: Applying existing artificial intelligence technologies to
the dogfight problem in experiments of increasing realism. In parallel, implementing
methods to measure, calibrate, increase, and predict human trust in combat autonomy
performance.
IMPACT: The program will scale the tactical application of autonomous dogfighting
to more complex, heterogeneous, multi-aircraft, operational-level simulated scenarios
informed by live data, laying the groundwork for future live, campaign-level Mosaic
Warfare experimentation.

Full Details: The ACE program seeks to increase trust in combat autonomy by
using human-machine collaborative dogfighting as its challenge problem. This also
serves as an entry point into complex human-machine collaboration. ACE will apply
existing artificial intelligence technologies to the dogfight problem in experiments of
increasing realism. In parallel, ACE will implement methods to measure, calibrate,
increase, and predict human trust in combat autonomy performance. Finally, the
program will scale the tactical application of autonomous dogfighting to more com-
plex, heterogeneous, multi-aircraft, operational-level simulated scenarios informed by
live data, laying the groundwork for future live, campaign-level Mosaic Warfare ex-
perimentation.In a future air domain contested by adversaries, a single human pilot
can increase lethality by effectively orchestrating multiple autonomous unmanned
platforms from within a manned aircraft. This shifts the human role from single
platform operator to mission commander. In particular, ACE aims to deliver a capa-
bility that enables a pilot to attend to a broader, more global air command mission
while their aircraft and teamed unmanned systems are engaged in individual tactics.
ACE creates a hierarchical framework for autonomy in which higher-level cognitive
functions (e.g., developing an overall engagement strategy, selecting and prioritizing
targets, determining best weapon or effect, etc.) may be performed by a human,
while lower-level functions (i.e., details of aircraft maneuver and engagement tactics)
is left to the autonomous system. In order for this to be possible, the pilot must
be able to trust the autonomy to conduct complex combat behaviors in scenarios
such as the within visual range dogfight before progressing to beyond visual range
engagements.The ACE demonstrations will bridge the gap from simple physics-based

13
automated systems currently in use to complex systems capable of effective autonomy
within highly dynamic and uncertain environments at mission speeds. The technol-
ogy development on the ACE program addresses four primary challenges: Increase
air combat autonomy performance in local behaviors (individual aircraft and team
tactical) Build and calibrate trust in air combat local behaviors Scale performance
and trust to global behaviors (heterogeneous multi-aircraft) Build infrastructure for
full-scale air combat experimentation

14
Network Universal Persistence (Net-
work UP)
Mr. Aaron Kofford

https://www.darpa.mil/program/network-universal-persistence.html

GPT-3 Summary: The goal of the Network UP program is to develop technol-

ogy that can maintain network reliability during periods of signal degradation. The
methods used will involve a new approach and architecture to improve performance.
If successful, this could have a major impact on communications in dynamic wireless
environments.

Full Details: The Network Universal Persistence (Network UP) program seeks to
develop and demonstrate radio technology that maintains network reliability through
periods of frequent signal degradation that may occur during operations in multiple
environments. From time to time, network outages may occur and data transmission
may be challenged. Networks in dynamic wireless environments can end up mostly
attempting to establish the network rather than sending data. The Network UP
program is addressing this issue by employing a new approach and novel architecture
to improve performance.

15
Guaranteed Architecture for Physi-
cal Security (GAPS)
Dr. Matthew Wilding

https://www.darpa.mil/program/guaranteed-architecture-for-physical-security.html

GPT-3 Summary: GOAL:

The GAPS program seeks to reduce inherent system complexity through the
development of hardware and software that is open, extendible, and compatible with
Size, Weight, and Power (SWaP) constrained environments to enable security across
DoD and commercial systems.

METHODS:

The goal of the GAPS program is to develop hardware security and software
architectures with provable security interfaces, which physically isolate high-risk
transactions during system design and system build, and track that such protections
are physically enforced at run-time.

IMPACT:

GAPS aims to substantially lower

Full Details: Modern computing systems are incapable of creating sufficient secu-
rity protections such that they can be trusted with the most sensitive data while
simultaneously being exposed to untrusted data streams. In certain places, the De-
partment of Defense (DoD) and commercial industry have adopted a series of air-gaps
– or breaks between computing systems – to prevent the leakage and compromise of
sensitive information. However, the use of air-gaps does not address the DoD’s need
for fusion of data across systems of different levels to support defense operations.
Similarly, the increasing use of cloud architectures demands better solutions for data
privacy in commercial computing systems.The Guaranteed Architectures for Physical
Security (GAPS) program seeks to reduce inherent system complexity through the
development of hardware and software that is open, extendible, and compatible with
Size, Weight, and Power (SWaP) constrained environments to enable security across
DoD and commercial systems. The goal of the GAPS program is to develop hardware
security and software architectures with provable security interfaces, which physically

16
isolate high-risk transactions during system design and system build, and track that
such protections are physically enforced at run-time. GAPS aims to substantially
lower the barrier to safely enable high-risk transactions, thus allowing for: a) faster
computer to computer transactions; b) spatial isolation reducing the need for insuffi-
cient software partitioning solutions (i.e. hypervisors); and c) more complex missions
without putting sensitive data at risk.

17
Coded Visibility
Dr. Rohith Chandrasekar

https://www.darpa.mil/program/coded-visibility.html

GPT-3 Summary: GOAL: Develop new obscurant systems that provide warfighters
with an asymmetric advantage by enhancing their visibility while suppressing adver-
sary visibility and detection.
TECHNOLOGY METHODS: Explore passive methods to achieve this asymmetry
using tailorable obscurants deployed in relevant environments. Focus on the longer-
term technical goal of actively modulating novel particulates, potentially enabling
dynamic control of an obscurant’s properties post-deployment and cooperation with
sensor modalities.
IMPACT: Improve the effectiveness of military tools while reducing health risks for
warfighters.

Full Details: Current obscurants are common military tools used to protect warfight-
ers from detection by adversary’s vision and sensors. However, despite decades of de-
velopment, obscurants have significant limitations: (1) they simultaneously degrade
the visual capability of friendly forces in addition to adversary forces; (2) once de-
ployed, their performance is fixed and cannot be tuned in real time; and (3) they pose
a serious health risk, requiring the use of respirators.The goal of the DARPA Coded
Visibility program is to address these limitations by developing next-generation ob-
scurant systems that provide warfighters an asymmetric advantage, enhancing their
visibility while suppressing adversary visibility and detection. Coded Visibility will
explore passive methods to achieve this asymmetry (i.e., passive asymmetry), us-
ing tailorable obscurants deployed in relevant environments. Additionally, Coded
Visibility will focus on the longer-term technical goal of actively modulating novel
particulates, potentially enabling dynamic control of an obscurant’s properties post-
deployment and cooperation with sensor modalities (i.e., active asymmetry). Across
both passive and active thrusts, the program will focus on identifying obscurants that
are safe for the warfighter.

18
Wideband Adaptive RF Protection
(WARP)
Dr. Timothy Hancock

https://www.darpa.mil/program/wideband-adaptive-rf-protection.html

GPT-3 Summary: GOAL: The WARP program will develop new methods to pro-
tect wideband RF receivers against interference.
TECHNOLOGY METHODS: Develop new filter architectures and signal cancellation
architectures that can adapt to the signals present in the spectrum.
IMPACT: Enable the use of wideband software-defined radios in contested and con-
gested spectral environments.

Full Details: Over the last decade, wideband analog-to-digital converter (ADC)
technology has improved in both bandwidth and resolution to a point that wideband
RF sampling receivers are now a reality. However, wideband ADCs typically have
less spur-free dynamic range as compared to their narrowband counterparts and are
typically exposed to more signals simultaneously due to the wide bandwidth. De-
spite the advantages associated with more bandwidth, the dynamic range limitation
can prevent the use of wideband receivers in multi-function applications that support
wideband electromagnetic spectrum operations (EMSO). The Wideband Adaptive
RF Protection (WARP) program seeks develop wideband adaptive filtering and signal
cancellation architectures to protect emerging wideband receivers against both exter-
nal and self-interference, ultimately enabling the use of wideband software-defined
radios in contested and congested spectral environments. To accomplish its goals,
WARP is focused on two key technical areas. In technical area 1, the focus is on
mitigating external interference in the 2-18 GHz band with tunable filtering that can
adapt to the signals present in the spectrum. This will be accomplished by first
targeting new filter architectures and associated underlying technology to achieve a
2:1 and 3:1 tuning ratio in Phase 1 and Phase 2 of the program, respectively, before
culminating in full-band coverage by the end of the program. In technical area 2,
the focus is on mitigating self-interference from a co-located transmitter in the 0.1-6
GHz band to enable same-frequency simultaneous transmit and receive (STAR). To
achieve wideband operation, this will require the canceller to have a time-bandwidth
product of approximately 10 by the end of the program, well beyond the state-of-the-
art in simple resonant RF canceller circuits. In these new approaches to wideband
reconfigurable filtering and signal cancellation, it is expected that the number of tun-
ing inputs could be large. To manage this tuning challenge, it is expected that the

19
hardware will adapt to the environment in real-time through embedded sensing and
control. If successful, by the end of the program, when exposed to external or self-
interference, WARP filters and cancellers will autonomously adapt to the spectral
environment to protect a wideband RF receiver.

20
High power Amplifier using Vacuum
electronics for Overmatch Capabil-
ity (HAVOC)

https://www.darpa.mil/program/high-power-amplifier-using-vacuum-electronics-for-overmatch-
capability.html

GPT-3 Summary: The goal of the HAVOC program is to develop mm-wave vacuum
electronic amplifiers to improve the performance of EM systems. The methods involve
developing new amplifiers and integrating them into EM systems. The impact is that
these systems will be able to access the high-frequency millimeter-wave portion of
the electromagnetic spectrum, resulting in increased range and other performance
improvements.

Full Details: The effectiveness of combat operations across all domains increasingly
depends on our ability to control and exploit the electromagnetic (EM) spectrum and
to deny its use to our adversaries. Below 30 GHz, the proliferation of inexpensive
high-power commercial radio frequency (RF) sources has made the EM spectrum
crowded and contested, challenging our spectrum dominance. The numerous tac-
tical advantages offered by operating at higher frequencies, most notably the wide
bandwidths available, is driving both commercial and DoD solid-state and vacuum
electronic amplifiers into the millimeter wave (mm-wave) spectrum above 30 GHz.
Control of the mm-wave spectrum requires advanced and ever more sophisticated
electronic components and systems. The performance of these systems strongly de-
pends on the available amplifier power.. The High power Amplifier using Vacuum
electronics for Overmatch Capability (HAVOC) program seeks to develop mm-wave
vacuum electronic amplifiers for air, ground, and ship-based EM systems. HAVOC
amplifiers would enable these systems to access the high-frequency millimeter-wave
portion of the electromagnetic spectrum, facilitating increased range and other per-
formance improvements such as high data-rate communications and high-resolution
sensing. Additionally, in the future the HAVOC program plans to support basic
research to improve the fundamental understanding of the various phenomena gov-
erning the science and technology that will underlie the next generation of mm-wave
vacuum electronic amplifiers.

21
Science of Artificial Intelligence and
Learning for Open-world Novelty (SAIL-
ON)
Mr. Ted Senator

https://www.darpa.mil/program/science-of-artificial-intelligence-and-learning-for-open-
world-novelty.html

GPT-3 Summary: GOAL: The Science of Artificial Intelligence and Learning for
Open-world Novelty (SAIL-ON) program will research and develop the underlying
scientific principles, general engineering techniques, and algorithms needed to create
AI systems that act appropriately and effectively in novel situations that occur in
open worlds.

TECHNOLOGY METHODS: The program’s goals are to develop scientific

principles to quantify and characterize novelty in open-world domains, create AI
systems that react to novelty in those domains, and demonstrate and evaluate these
systems in a selected DoD domain.

IMPACT: The goal of this program

Full Details: Current artificial intelligence (AI) systems excel at tasks defined by
rigid rules – such as mastering the board games Go and chess with proficiency sur-
passing world-class human players. However, AI systems aren’t very good at adapting
to constantly changing conditions commonly faced by troops in the real world – from
reacting to an adversary’s surprise actions, to fluctuating weather, to operating in
unfamiliar terrain. For AI systems to effectively partner with humans across a spec-
trum of military applications, intelligent machines need to graduate from closed-world
problem solving within confined boundaries to open-world challenges characterized
by fluid and novel situations. The Science of Artificial Intelligence and Learning for
Open-world Novelty (SAIL-ON) program intends to research and develop the under-
lying scientific principles, general engineering techniques, and algorithms needed to
create AI systems that act appropriately and effectively in novel situations that occur
in open worlds. The program’s goals are to develop scientific principles to quantify
and characterize novelty in open-world domains, create AI systems that react to nov-
elty in those domains, and demonstrate and evaluate these systems in a selected DoD
domain.

22
Resilient Anonymous Communica-
tion for Everyone (RACE)
Dr. Joshua Baron

https://www.darpa.mil/program/resilient-anonymous-communication-for-everyone.html

GPT-3 Summary: The RACE program’s goal is to create a more secure messaging
system that is resistant to large-scale compromise. The program will explore methods
for preventing information from being used to identify system nodes, and for obfus-
cating communication protocols. If successful, the program could have a significant
impact on the security of communications systems.

Full Details: The Resilient Anonymous Communication for Everyone (RACE) pro-
gram will research technologies for a distributed messaging system that can: a) exist
completely within a given network, b) provide confidentiality, integrity, and availabil-
ity of messaging, and c) preserve privacy to any participant in the system. Compro-
mised system data and associated networked communications should not be helpful
for compromising any additional parts of the system. RACE advances will be based
on rigorous security arguments, such as those found in the academic cryptography
community or statistical arguments based on realistic simulations. RACE will seek to
create advances in communication protocol encapsulation methods as well as efficient,
oblivious, distributed system tasking to build a system that is resistant to attack, even
with limited participant compromises and largescale, real-time deep packet inspec-
tion. The program will further seek to explore approaches to preserving privacy,
such as secure multiparty computation and obfuscated communication protocols.The
goal of the RACE program is to create a system capable of avoiding large-scale com-
promise. As such, RACE research efforts will explore: 1) preventing compromised
information from being useful for identifying any of the system nodes because all such
information is encrypted on the nodes at all times, even during computation; and 2)
preventing communications compromise by virtue of obfuscating communication pro-
tocols.Additional information is available in the RACE BAA.

23
Hierarchical Identify Verify Exploit
(HIVE)
Dr. Bryan Jacobs

https://www.darpa.mil/program/hierarchical-identify-verify-exploit.html

GPT-3 Summary: GOAL: The HIVE program seeks to develop a graph analytics
processor that can process streaming graphs 1000X faster and at much lower power
than current processing technology.
METHODS: The HIVE program will develop hardware and software to solve tough
challenges in areas such as cyber security and infrastructure monitoring.
IMPACT: The HIVE program will enable graph analytics techniques powerful enough
to solve tough challenges in areas such as cyber security and infrastructure monitoring.

Full Details: Social media, sensor feeds, and scientific studies generate large amounts
of valuable data. However, understanding the relationships among this data can be
challenging. Graph analytics has emerged as an approach by which analysts can
efficiently examine the structure of the large networks produced from these data
sources and draw conclusions from the observed patterns. By understanding the
complex relationships both within and between data sources, a more complete picture
of the analysis problem can be understood. With lessons learned from innovations in
the expanding realm of deep neural networks, the Hierarchical Identify Verify Exploit
(HIVE) program seeks to advance the arena of graph analytics. The HIVE program is
looking to build a graph analytics processor that can process streaming graphs 1000X
faster and at much lower power than current processing technology. If successful,
the program will enable graph analytics techniques powerful enough to solve tough
challenges in cyber security, infrastructure monitoring and other areas of national
interest. Graph analytic processing that currently requires racks of servers could
become practical in tactical situations to support front-line decision making. What
’s more, these advanced graph analytics servers could have the power to analyze the
billion- and trillion-edge graphs that will be generated by the Internet of Things, ever-
expanding social networks, and future sensor networks. In parallel with the hardware
development of a HIVE processor, DARPA is working with MIT Lincoln Laboratory
and Amazon Web Services (AWS) to host the HIVE Graph Challenge with the goal
of developing a trillion-edge dataset. This freely available dataset will spur innovative
software and hardware solutions in the broader graph analysis community that will
contribute to the HIVE program. The overall objective is to accelerate innovation in
graph analytics to open new pathways for meeting the challenge of understanding an

24
ever-increasing torrent of data. The HIVE program features two primary challenges:
The first is a static graph problem focused on sub-graph Isomorphism. This task is to
further the ability to search a large graph in order to identify a particular subsection
of that graph. The second is a dynamic graph problem focused on trying to find
optimal clusters of data within the graph. Both challenges will include a small graph
problem in the billions of nodes and a large graph problem in the trillions of nodes.

25
Direct On-Chip Digital Optical Syn-
thesizer (DODOS)
Dr. Gordon Keeler

https://www.darpa.mil/program/direct-on-chip-digital-optical-synthesizer.html

GPT-3 Summary: GOAL: The Direct On-Chip Digital Optical Synthesizer (DO-
DOS) program seeks to create a technology revolution in optical frequency control.
TECHNOLOGY METHODS: The program proposes to use recent breakthroughs in
chip-scale mode-locked lasers and microresonators to enable self-referenced optical
frequency combs in compact integrated packages.
IMPACT: The DODOS chip will act as a gearbox to directly translate the stabil-
ity and accuracy of an RF signal to the optical domain, producing laser light at
frequencies above 200 terahertz with relative accuracy of one

Full Details: The Direct On-Chip Digital Optical Synthesizer (DODOS) program
seeks to create a technological revolution in optical frequency control analogous to
the disruptive advances in microwave frequency control in the 1940s. That early
development ushered in a new era for microwave technology, transformed modern
warfare, and has since been enabling a multitude of Department of Defense (DoD)
and civilian capabilities, including radar, navigation technologies, and satellite and
terrestrial communications. Extending frequency control to the optical regime is an-
ticipated to greatly extend the technology base for the next generation of warfighter
and other capabilities. Since the first demonstration of optical frequency synthesis
using self-referenced optical combs in 2000, demonstrations of novel civilian and de-
fense applications for the technology have emerged worldwide. Due to the large size,
relative fragility, and high cost of these components and systems, however, precise op-
tical frequency synthesis has been limited to lab-scale experiments. DODOS aims to
leverage recent breakthroughs in chip-scale mode-locked lasers and microresonators to
enable self-referenced optical frequency combs in compact integrated packages. The
program proposes to draw on recent progress in heterogeneous photonic integration
to arrange all of the necessary components, including widely tunable laser sources,
optical modulators, non-linear photonic elements, and CMOS radiofrequency (RF)
and control circuitry onto a chip with a volume of less than 1 cmˆ3 and power con-
sumption of less than 1 Watt. The DODOS chip will act as a gearbox to directly
translate the stability and accuracy of an RF signal to the optical domain, producing
laser light at frequencies above 200 terahertz with relative accuracy of one part in
10ˆ15. New generations of optical frequency control technology could enable a wide

26
range of applications in optical spectroscopy, gas sensing, LiDAR, portable atomic
clocks, high-bandwidth and secure communications, and intrusion detection, among
other areas. The DODOS program also aims to spur research in octave-spanning mi-
crocombs, high-efficiency chip-scale lasers, high-efficiency on-chip frequency doubling,
and CMOS-compatible integration technology.

27
No Manning Required Ship (NO-
MARS)
Dr. Gregory Avicola

https://www.darpa.mil/program/no-manning-required-ship.html

GPT-3 Summary: The No Manning Required Ship (NOMARS) program seeks to

design a ship that can operate autonomously for long durations at sea. NOMARS
focuses on exploring novel approaches to the design of the seaframe (the ship with-
out mission systems) while accommodating representative payload size, weight, and
power. NOMARS aims to challenge the traditional naval architecture model, de-
signing a seaframe from the ground up with no provision, allowance, or expectation
for humans at sea. By removing the human element from all ship design consider-
ations, the program intends to demonstrate significant advantages, to include size,
cost, at-sea

Full Details: The No Manning Required Ship (NOMARS) program seeks to design
a ship that can operate autonomously for long durations at sea, enabling a clean-sheet
ship design process that eliminates design considerations associated with crew. NO-
MARS focuses on exploring novel approaches to the design of the seaframe (the ship
without mission systems) while accommodating representative payload size, weight,
and power. NOMARS aims to challenge the traditional naval architecture model, de-
signing a seaframe from the ground up with no provision, allowance, or expectation
for humans at sea. By removing the human element from all ship design consider-
ations, the program intends to demonstrate significant advantages, to include size,
cost, at-sea reliability, survivability to sea-state, and survivability to adversary actions
such as stealth considerations and resistance to tampering. The program also will
strive for greater hydrodynamic efficiency via hull optimization without requirements
for crew safety or comfort.

28
All Together Now (ATN)
Dr. Tatjana Curcic

https://www.darpa.mil/program/all-together-now.html

GPT-3 Summary: The goal of All Together Now (ATN) is to develop theoreti-
cal protocols and experimental techniques that enable new collective atom regimes,
leading to sensitivities approaching the ultimate fundamental limits of performance.
ATN investigates superradiant lasing, a lasing regime that may lead to unprecedented
laser linewidths and coherence, that is insensitive to environmental perturbation. The
program also explores the potential for novel collective atomic laser cooling regimes,
and aims to develop an entangled 3D optical lattice clock. If successful, this sys-
tem could yield environmentally insensitive ultra-narrow linewidth lasers and optical
atomic clocks with one

Full Details: The goal of All Together Now (ATN) is to develop theoretical proto-
cols and experimental techniques that enable new collective atom regimes, leading to
sensitivities approaching the ultimate fundamental limits of performance. Quantum
sensors, such as atomic clocks and atom interferometers, are currently reaching the
independent atom limit in which the uncertainty of the sensor scales with the number
√
N of atoms as 1/ N (known as the standard quantum limit). However, the ultimate
fundamental limit of performance scales as 1/N if one allows correlations between
the atoms. In quantum sensors such as optical lattice clocks using trapped reservoirs
of tens of thousands of atoms, improvements of a few orders of magnitude are thus
possible. In order to approach this fundamental limit, however, the system must
leverage collective quantum effects that have until now proven difficult to prepare
and observe. ATN investigates superradiant lasing, a lasing regime that may lead
to unprecedented laser linewidths and coherence, that is insensitive to environmental
perturbation. The program also explores the potential for novel collective atomic
laser cooling regimes, and aims to develop an entangled 3D optical lattice clock. If
successful, this system could yield environmentally insensitive ultra-narrow linewidth
lasers and optical atomic clocks with one-second stabilities exceeding current state-
of-the-art flicker floors reached after thousands of seconds of averaging times. ATN
developments will not only provide a feasibility study of atomic clock operation at
the fundamental limit, but the techniques developed in this program may be inte-
grated into the next generation of timekeeping devices, IMUs, magnetometers, and
gravimeters.

29
Fieldable Solutions for Hemorrhage
with bio-Artificial Resuscitation Prod-
ucts (FSHARP)
CDR Jean-Paul Chretien

https://www.darpa.mil/program/fieldable-solutions-for-hemorrhage-with-bio-artificial-
resuscitation-products.html

GPT-3 Summary: The FSHARP program will develop a whole blood substitute as
a hemorrhage countermeasure to sustain injured warfighters in austere, pre-hospital
settings. The program will deliver formulations that recapitulate the most critical
functions of whole blood without the requirement for cold chain storage. These
formulations will be tailorable to different injury profiles, making them suitable for
use in active combat zones, disaster response, and medical settings worldwide.

Full Details: In both military and civilian settings, hemorrhage is the leading cause
of survivable, traumatic pre-hospital death. Whole blood is considered the optimal
resuscitation fluid for traumatic hemorrhage in tactical combat casualty care, but the
cold-chain requirement causes logistical difficulties for far-forward units. To address
this challenge, the Fieldable Solutions for Hemorrhage with bio-Artificial Resuscita-
tion Products (FSHARP) program will develop a deployable, shelf-stable, universal
whole blood substitute as a hemorrhage countermeasure to sustain injured warfight-
ers in austere, pre-hospital settings. At the end of the program, deliverables will
be formulations that recapitulate the most critical functions of whole blood without
requirement for cold chain, and are tailorable to different injury profiles requiring
specific therapeutic adaptations (e.g., hemorrhage with trauma induced coagulopa-
thy, traumatic brain injury, or other concurrent injury). FSHARP formulations will
be suitable for use in active combat zones, disaster response, and medical settings
worldwide.

30
Measuring the Information Control
Environment (MICE)
Dr. Joshua Baron

https://www.darpa.mil/program/measuring-the-information-control-environment.html

GPT-3 Summary: GOAL: To develop technology that will automatically update

and feed into easily-understood dashboards that provide real-time ground truth un-
derstanding of how countries conduct domestic information control.
TECHNOLOGY METHODS: Algorithms and prototype open-source software that
measures how authoritarian regimes employ digitally repressive techniques, such as
censorship, blocking, or throttling, at scale over the internet, as well as the technical
capabilities used to enable such activities.
IMPACT: A better understanding of how countries censor, block, or throttle specific
internet-based activities and the technical capabilities that countries use to enable
such

Full Details: The Measuring the Information Control Environment (MICE) AI Ex-
ploration Opportunity aims to develop algorithms and prototype open-source software
that measures how authoritarian regimes employ digitally repressive techniques, such
as censorship, blocking, or throttling, at scale over the internet, as well as the technical
capabilities used to enable such activities. The goal of MICE is to develop technol-
ogy that will continuously and automatically update and feed into easily-understood
dashboards that provide comprehensive, real-time ground truth understanding of how
countries conduct domestic information control. MICE will measure how countries
censor, block, or throttle specific internet-based activities. MICE will also try to de-
termine the technical capabilities that countries use to enable such repressive activi-
ties. To comprehensively understand the scope of these activities, MICE technologies
will aim to discover and track changes over time to adversary techniques across vari-
ous abstractions of the information environment from the network level (e.g., netflow,
DNS, IP address, protocol, applications, etc.) to the information level (e.g., topics,
images, audio, video, etc.).Additional information is available in the MICE Program
Announcement.

31
Computational Cultural Understand-
ing (CCU)
Dr. William Corvey

https://www.darpa.mil/program/computational-cultural-understanding.html

GPT-3 Summary: GOAL: The program’s goal is to create cross-cultural language

understanding technologies to improve a DoD operator’s situational awareness and
interactional effectiveness.
TECHNOLOGY METHODS: The program will explore technologies engineered to
require minimal-to-no training data in a local culture, while maximizing operator
success during negotiations and other interactions in the field.
IMPACT: The technologies developed through this program will assist negotiators
and analysts with language analysis and cross-cultural dialogue in the field.

Full Details: The U.S. Government and its agencies, including the Department of
Defense (DoD), operate globally and are in constant contact with diverse cultures.
Communicative understanding, not simply of local languages, but also of social cus-
toms and cultural backgrounds, lies at the heart of Civil Affairs and Military In-
formation Support Operations activities, which together comprise the vast majority
of U.S. counterinsurgency and stabilization efforts. Cross-cultural miscommunica-
tion not only derails negotiations, but also can be a contributing factor leading to
war. The likelihood of communicative failure increases dramatically where signif-
icant social, cultural, or ideological differences exist.Automated systems would be
a welcome force-multiplier for DoD interpreters. However, unlike the human cul-
tural interpreters who enable U.S. forces today, such systems are currently incapable
of accurately analyzing cross-cultural communication or providing useful assistance
to negotiations beyond basic machine translation.The Computational Cultural Un-
derstanding (CCU) program aims to create cross-cultural language understanding
technologies to improve a DoD operator’s situational awareness and interactional
effectiveness. CCU seeks to build natural language processing technologies that rec-
ognize, adapt to, and recommend how to operate within the emotional, social, and
cultural norms that differ across societies, languages, and group affinities. To sup-
port diverse and emergent use cases, CCU will explore technologies engineered to
require minimal-to-no training data in a local culture, while maximizing operator
success during negotiations and other interactions in the field. Instead of relying
primarily on annotated training data, target systems will leverage qualitative and
quantitative findings from fields such as psychology, sociology, or other relevant disci-

32
plines, as well as minimally-supervised machine learning techniques, in order to infer
the meaning of unlabeled discourse behaviors in context. To achieve relative parity
with human interpreters, who provide cultural insight and leverage this knowledge in
the translation process, CCU research aims to provide foundational technical inno-
vations to assist negotiators and analysts with language analysis and cross-cultural
dialogue in the field. These advances in the area of cross-cultural understanding will
also inform processes for training and planning.Despite remarkable recent advances
in machine learning and multimedia analysis, today’s human language technology ap-
plications are inadequate for cross-cultural communicative analysis. To remedy these
deficiencies and advance communication technology towards enablement of greater
cultural understanding, the CCU program will focus on the following research top-
ics:Automatic discovery of sociocultural norms.Generalization of emotion recognition
across cultures.Detecting impactful changes in communicative practice at multiple
timescales. Providing dialogue assistance to cross-cultural interaction.For more infor-
mation CCU Broad Agency AnnouncementCCU FAQCCU Proposers Day SlidesCCU
Amendment 2CCU Amendment 1

33
Safe Genes
Dr. Anne Cheever

https://www.darpa.mil/program/safe-genes.html

GPT-3 Summary: The goal of the Safe Genes program is to create tools and
methodologies to control, counter, and even reverse the effects of genome edit-
ing—including gene drives—in biological systems across scales. The program is devel-
oping the genetic circuitry and genome editing machinery for robust, spatial, tempo-
ral, and reversible control of genome editing activity in living systems. Additionally,
the program is developing small molecules and molecular strategies to prevent or
limit genome editing activity and protect the genome integrity of organisms and pop-
ulations. Finally, the program is developing “genetic remediation” strategies that
eliminate unwanted engineered genes from a broad range of complex population

Full Details: The Safe Genes program supports force protection and military health
and readiness by protecting Service members from accidental or intentional misuse
of genome editing technologies. Additional work will leverage advances in gene edit-
ing technology to expedite development of advanced prophylactic and therapeutic
treatments against gene editors. Advances within the program will ensure the United
States remains at the vanguard of the broadly accessible and rapidly progressing
field of genome editing. Safe Genes performer teams work across three primary tech-
nical focus areas to develop tools and methodologies to control, counter, and even
reverse the effects of genome editing—including gene drives—in biological systems
across scales. First, researchers are developing the genetic circuitry and genome edit-
ing machinery for robust, spatial, temporal, and reversible control of genome editing
activity in living systems. Second, researchers are developing small molecules and
molecular strategies to provide prophylactic and treatment solutions that prevent
or limit genome editing activity and protect the genome integrity of organisms and
populations. Third, researchers are developing “genetic remediation” strategies that
eliminate unwanted engineered genes from a broad range of complex population and
environmental contexts to restore systems to functional and genetic baseline states.
Overall, the Safe Genes program is creating a layered, modular, and adaptable so-
lution set to: protect warfighters and the homeland against intentional or acciden-
tal misuse of genome editing technologies; prevent and/or reverse unwanted genetic
changes in a given biological system; and facilitate the development of safe, precise,
and effective medical treatments that use gene editors. Safety is a priority for the
program. All work performed under the program will occur in controlled, biosecure
facilities. Additionally, the Safe Genes program is informed by independent Ethical,

34
Legal, and Social Implications (ELSI) experts to help DARPA proactively identify
potential issues related to genome editing technologies. Performer teams will also
engage with potential stakeholders, including government regulators, to increase the
value of the science and to shape experiments around their questions and concerns.
These communications supplement the standard oversight provided by institutional
review boards that govern animal use and biosafety containment.

35
Fast Event-based Neuromorphic Cam-
era and Electronics (FENCE)
Dr. Whitney Mason

https://www.darpa.mil/program/fast-event-based-neuromorphic-camera-and-electronics.html

GPT-3 Summary: The FENCE program seeks to develop an integrated event-

based infrared (IR) FPA with embedded processing to overcome the challenges faced
by previous event-based cameras. The FENCE FPA could solve the problems of
timing accuracy and data sparsity to produce an infrared event-based neuromorphic
imager consistent with military requirements. The FENCE FPA will consist of a
low latency, high precision event-based readout matched to low power embedded
neuromorphic processing capable of executing algorithms that use combined spatial
and temporal (spatio-temporal) information, to produce a new intelligent sensor for
tactical DoD applications.

Full Details: Event-based imagers are an emerging class of sensors with signifi-
cant demonstrated advantages relative to traditional cameras. Because they operate
asynchronously and only transmit data from pixels that have changed, they have
been shown to produce over 100x less data in sparse scenes relative to traditional
focal plane arrays (FPAs). This leads directly to 100x lower latency at 100x lower
power. Despite their inherent advantages, existing event-based cameras are not cur-
rently compatible with Department of Defense (DoD) applications as DoD scenarios
are highly cluttered and dynamic. The Fast Event-based Neuromorphic Camera and
Electronics (FENCE) program seeks to develop an integrated event-based infrared
(IR) FPA with embedded processing to overcome the challenges faced by previous
event-based cameras. The FENCE FPA could solve the problems of timing accuracy
and data sparsity to produce an infrared event-based neuromorphic imager consis-
tent with military requirements. The FENCE FPA will consist of a low latency,
high precision event-based readout matched to low power embedded neuromorphic
processing capable of executing algorithms that use combined spatial and tempo-
ral (spatio-temporal) information, to produce a new intelligent sensor for tactical
DoD applications. The sole technical area is focused on the development of an asyn-
chronous read-out integrated circuit (ROIC) capable of very low latency and power
operation and a low power processing layer that integrates with the ROIC to identify
salient spatio-temporal signals. The goal is to integrate the ROIC and the processing
layer together to demonstrate a FENCE sensor operating at low power (< 1.5 W).

36
Target Recognition and Adaption in
Contested Environments (TRACE)
Dr. Adam Bryant

https://www.darpa.mil/program/trace.html

GPT-3 Summary: GOAL: To develop an accurate target recognition system that

can be used in conjunction with a radar for long-range tactical applications.

METHODS: Develop a system that can be co-located with a radar to provide

responsive target recognition in real-time.

IMPACT: Improve the accuracy of target recognition for airborne surveillance

and strike operations.

Full Details: In a target-dense environment, the adversary has the advantage of us-
ing sophisticated decoys and background traffic to degrade the effectiveness of existing
automatic target recognition (ATR) solutions. Airborne strike operations against re-
locatable targets require that pilots fly close enough to obtain confirmatory visual
identification before weapon release, putting the manned platform at extreme risk.
Radar provides a means for imaging ground targets at safer and far greater standoff
distances; but the false-alarm rate of both human and machine-based radar image
recognition is unacceptably high. Existing ATR algorithms also require impractically
large computing resources for airborne applications. Hence, current approaches for
inserting ATR into tactical applications either move the processing to remote ground
stations or drastically reduce performance to fit legacy airborne platform comput-
ing capabilities. The Target Recognition and Adaption in Contested Environments
(TRACE) program seeks to develop an accurate, real-time, low-power target recog-
nition system that can be co-located with the radar to provide responsive long-range
targeting for tactical airborne surveillance and strike applications.

37
Materials for Transduction (MATRIX)
Dr. William Carter

https://www.darpa.mil/program/matrix.html

GPT-3 Summary: GOAL: The program aims to develop materials that can effi-
ciently convert energy between different forms or domains.
METHODS: The program will use diverse modeling, design and fabrication commu-
nities in a unified research and development effort addressing applications that bridge
the material and the device domains.
IMPACT: The program could deliver materials, devices, and modeling tools that
can enable new transduction capabilities with significantly higher performance; lower
noise; and smaller size, weight and power than current state-of-the-art technologies.

Full Details: Transductional materials convert energy between different forms or

domains, such as thermal to electrical energy, or electric field to magnetic field. De-
vices fabricated from such materials have multiple DoD-relevant applications that
include the following: Thermoelectrics (thermal/electric domains) used for energy
harvesting, thermal management, and refrigeration; Multiferroics (magnetic/electric
domains) used in sensors, antennas, actuators, micromotors, tunable RF and mi-
crowave components; Phase Change Materials (various domains) used in transducers,
switches, sensors, and control devices. While significant progress has been made in
advancing transductional material performance for certain applications, gains at the
material level have not always translated into new devices and DoD capabilities. The
goal of MATRIX is to extend materials breakthroughs to the device and systems
level by integrating diverse modeling, design and fabrication communities in a uni-
fied research and development effort addressing applications that bridge the material
and the device domains. A major program thrust is the development of multiscale,
multimodal design and engineering tools that have the potential to accelerate adop-
tion of MATRIX technology into DoD platforms. Anticipated deliverables from the
MATRIX program include materials, devices, and modeling tools that can enable
new transduction capabilities with significantly higher performance; lower noise; and
smaller size, weight and power than current state-of-the-art technologies.

38
Reverse Engineering of Deceptions
(RED)
Dr. Bruce Draper

https://www.darpa.mil/program/reverse-engineering-of-deceptions.html

GPT-3 Summary: The RED program seeks to automatically reverse engineer the
toolchains behind information-based attacks, such as multimedia falsification, adver-
sarial ML attacks, or other deception-based attacks. The goal is to develop techniques
that support the automated identification of attack toolchains as well as the develop-
ment and maintenance of scalable databases of attack toolchains. This information
may aid in identifying an adversary and understanding the methods used in an attack.

Full Details: Machine Learning (ML) techniques are susceptible to adversarial de-
ception at training time and when deployed. Similarly, humans are susceptible to
being deceived by falsified media (images, video, audio, text) or other information
created with malicious intent. The consequences may be significant in both cases,
and deception plays an increasingly central role in information-based attacks. The
Reverse Engineering of Deceptions (RED) effort aims to develop techniques that
automatically reverse engineer the toolchains behind attacks, such as multimedia fal-
sification, adversarial ML attacks, or other information deception attacks. The tools
by which those attacks are accomplished and the adversaries behind such attacks are
often unclear. Recovering the tools and processes used to create an attack provides
information that may aid in identifying an adversary. RED seeks to develop tech-
niques that support the automated identification of attack toolchains as well as the
development and maintenance of scalable databases of attack toolchains.For more
information, see the RED Program Announcement.

39
Applications Resulting from Recent
Insights in Vacuum Engineering (AR-
RIVE)
Lt. Col. C. David Lewis, USAF

https://www.darpa.mil/program/applications-resulting-from-recent-insights-in-vacuum-
engineering.html

GPT-3 Summary: The goal of the ARRIVE program is to experimentally test

new insights into controlling the quantum vacuum, in order to develop improved
quantum systems. The program is focused on three main areas: the Casimir force,
the dynamical Casimir force, and spin qubit decoherence. If successful, the program
could lead to advances in nano-scale actuators, qubit gates for quantum computing,
and high-brightness entangled photon sources.

Full Details: The Applications Resulting from Recent Insights in Vacuum Engi-
neering (ARRIVE) program is carrying out proof-of-theory experiments aimed at
engineering vacuum fluctuations in photonic and mechanical systems. A follow-on
to the DARPA QUEST program, ARRIVE will experimentally test new theoretical
insights into controlling the quantum vacuum, and transducing the vacuum energy
density into a usable form. ARRIVE is broadly focused on three vacuum effects: the
Casimir force, the dynamical Casimir force, and spin qubit decoherence. If successful,
these focus areas will demonstrate that the quantum vacuum can be engineered in
ways that produce meaningful impacts that enable improved and unprecedented func-
tionality to quantum systems such as nano-scale actuators, qubit gates for quantum
computing, and high-brightness entangled photon sources.

40
Data-Driven Discovery of Models (D3M)
Dr. Joshua Elliott

https://www.darpa.mil/program/data-driven-discovery-of-models.html

GPT-3 Summary: The goal of the D3M program is to develop automated model
discovery systems that will enable users with subject matter expertise but no data
science background to create empirical models of real, complex processes. The three
key technologies that will be developed in order to achieve this goal are: a library of
selectable primitives, automated composition of complex models, and human-model
interaction that enables curation of models by subject matter experts. The auto-
mated model discovery systems developed by the D3M program will be tested on
real-world problems that will progressively get harder during the course of the pro-
gram. Ultimately, the goal is to target

Full Details: Understanding the complex and increasingly data-intensive world

around us relies on the construction of robust empirical models, i.e., representations
of real, complex systems that enable decision makers to predict behaviors and answer
“what-if” questions. Today, construction of complex empirical models is largely a
manual process requiring a team of subject matter experts and data scientists. With
ever more data becoming available via improved sensing and open sources, the op-
portunity exists to build models to speed scientific discovery, enhance Department
of Defense/Intelligence Community’s intelligence, and improve United States Gov-
ernment logistics and workforce management, but capitalizing on this opportunity is
fundamentally limited by the availability of data scientists. The Data-Driven Discov-
ery of Models (D3M) program aims to develop automated model discovery systems
that enable users with subject matter expertise but no data science background to
create empirical models of real, complex processes. This capability will enable sub-
ject matter experts to create empirical models without the need for data scientists,
and will increase the productivity of expert data scientists via automation. The D3M
automated model discovery process, depicted in the figure, will be enabled by three
key technologies to be developed in the course of the program: A library of selectable
primitives. A discoverable archive of data modeling primitives will be developed
to serve as the basic building blocks for complex modeling pipelines. Automated
composition of complex models. Techniques will be developed for automatically se-
lecting model primitives and for composing selected primitives into complex modeling
pipelines based on user-specified data and outcome(s) of interest. Human-model in-
teraction that enables curation of models by subject matter experts. A method and
interface will be developed to facilitate human-model interaction that enables formal

41
definition of modeling problems and curation of automatically constructed models
by users who are not data scientists. Automated model discovery systems developed
by the D3M program will be tested on real-world problems that will progressively
get harder during the course of the program. Toward the end of the program, D3M
will target problems that are both unsolved and underspecified in terms of data and
instances of outcomes available for modeling.

42
Biological Control
Dr. Paul Sheehan

https://www.darpa.mil/program/biological-control.html

GPT-3 Summary: GOAL:

The goal of the Biological Control program is to establish design and control
principles that lead to reliable performance in biological systems.

METHODS:

The methods used in this program include developing technologies that enable
consistent operation of systems that combat biological threats, speed healing after
physical trauma, and support military readiness by complementing the body’s
natural defenses against emerging diseases.

IMPACT:

The impact of this program will be to enable a wide range of potential appli-
cations for the Department of Defense, including combatting biological threats,
speeding healing, and supporting military readiness.

Full Details: The Biological Control program seeks to support a wide range of po-
tential Department of Defense (DoD) applications by establishing design and control
principles that lead to reliable performance in biological systems. Leveraging tech-
nologies developed under this program will enable consistent operation of systems that
combat biological threats; speed healing after physical trauma; and support military
readiness by complementing the body’s natural defenses against emerging diseases.

43
Control of Revolutionary Aircraft with
Novel Effectors (CRANE)
Dr. Alexander M. G. Walan

https://www.darpa.mil/program/control-of-revolutionary-aircraft-with-novel-effectors.html

GPT-3 Summary: GOAL: The CRANE program seeks to develop a novel X-plane
that uses Active Flow Control (AFC) to improve aircraft performance.
TECHNOLOGY METHODS: The program will mature AFC technologies and design
tools, and incorporate them early in the design process.
IMPACT: The demonstration of AFC could help open the design trade space for
future military and commercial applications.

Full Details: The Control of Revolutionary Aircraft with Novel Effectors (CRANE)
program aims to design, build, and flight test a novel X-plane that incorporates Ac-
tive Flow Control (AFC) as a primary design consideration. Crane seeks to optimize
the benefits of active flow control by maturing technologies and design tools, and
incorporating them early in the design process. Active flow control could improve
aircraft performance by removing jointed surfaces, which currently drive design con-
figurations that increase weight and mechanical complexity. Demonstrating AFC
for stability and control in-flight would help open the design trade space for future
military and commercial applications.

44
Generating RF with Photonic Oscil-
lators for Low Noise (GRYPHON)
Dr. Gordon Keeler

https://www.darpa.mil/program/generating-rf-with-photonic-oscillators-for-low-noise.html

GPT-3 Summary: The GRYPHON program seeks to develop extremely low phase
noise microwave frequency oscillators in a small, compact package. This would enable
advanced sensing and communication applications. The goal is to have these devices
meet or exceed the performance of the best discrete oscillator modules, yet occupy a
much smaller volume.

Full Details: Electronic oscillators lie at the heart of virtually all microelectronic sys-
tems, generating the periodic clock signals used in digital electronics and the precise
frequencies that enable radar and radio frequency (RF) communications. While an
ideal oscillator provides a tone at a singular frequency, component imperfections and
coupling to the environment degrade the spectral purity of real-world sources. These
impairments, broadly quantified as phase noise, ultimately limit the performance of
many military and commercial RF systems.The Generating RF with Photonic Os-
cillators for Low Noise (GRYPHON) program seeks to develop compact microwave
frequency oscillators with extremely low phase noise to enable advanced sensing and
communication applications. In the last decade, major advances in oscillator per-
formance have been realized using optical techniques to synthesize high-fidelity mi-
crowave signals. Such oscillators employ optical frequency division to reach world-
record phase noise levels. The solutions demonstrated to date, however, have sac-
rificed other important attributes in pursuit of spectral purity. Such trade-offs are
problematic, because module size, cost, tunability, and environmental sensitivity are
also critical factors that determine the applicability of microwave sources to commer-
cial and military systems. GRYPHON will leverage recent developments in nonlinear
photonics and photonic-electronic integration to develop microwave sources with noise
performance that meets or exceeds that of the best discrete oscillator modules, yet
occupy a compact volume typical of far noisier chip-scale voltage-controlled oscilla-
tors (VCOs). Moreover, by program end, GRYPHON microwave sources will operate
as synthesizers with the ability to tune to any frequency from 1 to 40+ GHz during
operation. This combination of features is unprecedented in today’s state of the art,
and will establish a new regime of source technology that is expected to transform the
types and capabilities of military and commercial radar and communication systems.

45
Joint University Microelectronics Pro-
gram (JUMP)
Dr. Dev Palmer

https://www.darpa.mil/program/joint-university-microelectronics-program.html

GPT-3 Summary: The goal of the JUMP program is to improve microelectronic

technologies in order to enable advances in cognitive computing, intelligent memory
and storage, distributed computing and networking, and radio frequency (RF) to ter-
ahertz (THz) sensor and communications systems. The technology methods used in
this program include developing new architectures and algorithms, and advanced de-
vices, packaging, and materials. The impact of this program will be the development
of systems that are transferable to military and industry in a five year timeframe and
ready for field deployment in 10.

Full Details: Due to engineering limitations and cost constraints, the dynamics of
the electronic industry are continually changing. Commercial companies increasingly
recognize the need to differentiate their products through research in areas other than
device scaling, such as new circuit architectures and computing algorithms. At the
same time, the DoD has increasing computation needs in autonomous and distributed
systems that will require advances beyond the current Moore’s Law roadmap. Cog-
nitive electronics warfare (EW), secure communications, and advanced imaging will
all need computation performance at power that will require the most advanced fab-
rication processes, but also new architectures and computing algorithms. Without
these advances, the promise of new algorithms and capabilities for the warfighter will
not be realized. DARPA, along with companies from the semiconductor and defense
industries—Intel, IBM, Micron, Analog Devices, EMD Performance Materials, ARM,
Samsung, TSMC, Raytheon, Northrop Grumman, and Lockheed Martin—have ini-
tiated the Joint University Microelectronics Program (JUMP) with six research cen-
ters to undertake high-risk, high-payoff research that addresses existing and emerging
challenges in microelectronic technologies. JUMP comes at an inflection point in the
history of the semiconductor industry where application and system research is criti-
cal to enabling the development of superior electronic systems to meet both DoD and
commercial needs. Under JUMP, the challenges of the “application-centric” research
centers focus on accomplishing application-oriented goals and spurring the develop-
ment of complex systems with capabilities well beyond those available today. Diving
deep into cognitive computing, intelligent memory and storage, distributed computing
and networking, and radio frequency (RF) to terahertz (THz) sensor and communica-

46
tions systems, among other areas, these research centers will strive to develop systems
that will be transferable to military and industry in a five year timeframe and ready
for field deployment in 10. The four application-centric JUMP centers are: The Cen-
ter for Brain-inspired Computing Enabling Autonomous Intelligence (C-BRIC) aims
to deliver major advances in cognitive computing, with the goal of enabling a new
generation of autonomous intelligent systems. The Center for Converged TeraHertz
Communications and Sensing (ComSecTer) seeks to develop technologies for a future
cellular infrastructure designed to support the autonomous vehicle revolution and
the emergence of intelligent highways. The Computing On Network Infrastructure
for Pervasive Perception, Cognition, and Action (CONIX) center is striving to de-
velop an architecture for networked computing that lies between edge devices and
the cloud. The Center for Research on Intelligent Storage and Processing-in-memory
(CRISP) is working to topple the “memory wall”–a 70-year-old technical bottleneck
in computer systems that is hindering the use of big data for technical discovery.
In addition, two “disciplinary” research centers will take on the challenge of driv-
ing foundational developments around specific disciplines with the goal of creating
disruptive breakthroughs in areas relevant to JUMP sponsors, including advanced
architectures and algorithms, and advanced devices, packaging, and materials. The
two disciplinary JUMP centers are: The Applications Driving Architectures (ADA)
Center is researching new areas that reduce the cost, complexity, and energy required
to develop advanced computing systems by democratizing the design and manufac-
turing process. The Applications and Systems driven Center for Energy-Efficient
Integrated Nanotechnologies (ASCENT) focuses on material and device innovations
to transcend the anticipated limits of current CMOS technology in order to increase
the performance, efficiency, and capabilities of future computing systems.

47
Cornucopia
Dr. Molly Jahn

https://www.darpa.mil/program/cornucopia.html

GPT-3 Summary: The Cornucopia program seeks to enable the on-demand pro-
duction of microbial-origin food starting from water, air, and electricity. If successful,
the program could significantly reduce the logistical burden of food transportation as-
sociated with deployed military operations and humanitarian assistance and disaster
relief (HADR) operations.

Full Details: Cornucopia seeks to enable deployable, on-demand production of appe-

tizing, microbial-origin food starting from water, air, and electricity with minimal to
no supplementation. If successful, the program could significantly reduce the logistical
burden of food transportation associated with deployed military operations and hu-
manitarian assistance and disaster relief (HADR) operations. To address vulnerabili-
ties in food supply chains across a variety of operational and humanitarian scenarios,
Cornucopia will demonstrate the capacity to produce all four human dietary macronu-
trients (protein, carbohydrate, fat, and dietary fiber) in ratios that target Military
Dietary Reference Intake (MDRI) daily requirements for complete nutrition. Out-
puts will be in multiple food formats (e.g., shake, bar, gel, jerky) that meet military
nutritional standards and palatability requirements in a system minimizing inputs,
handling, and footprint.Cornucopia is a four-year program with three focus areas:
domestication of microbes for human consumption; tailorability of microbial-origin
food for different flavors, formats and nutritional composition; and integrated system
demonstrations benchmarked by two military use cases, a small forward-operating
military unit in austere conditions and a HADR scenario.Proposers Day VideoMr.
Peter Donaghue, DARPA Contracts Management OfficeHow to propose to the Cor-
nucopia program: 0:00 – 11:17Dr. Phil Root, Acting Director, DARPA Defense Sci-
ences Office (DSO)DARPA overview and unique role of the Defense Sciences Office
in revolutionary RandD: 11:18 – 22:29Dr. Molly Jahn, Cornucopia Program Manager
in DSODetails and vision for the Cornucopia program: 22:30 – 58:03

48
LogX
Dr. Jeff Maas

https://www.darpa.mil/program/logx.html

GPT-3 Summary: The LogX program’s goal is to develop software for real-time
logistics and supply chain system situational awareness (diagnosis), future state pre-
diction (prognosis), and assessment of resilience at unprecedented scale and speed.
The program’s technology methods include developing software to exploit the recent
migration of logistics information to digital formats and the cloud. The program’s
impact is to build a capability to work alongside existing logistics information sys-
tems to improve the Department of Defense’s ability to manage the flow of millions
of military parts, supplies, and pieces of equipment around the world.

Full Details: The Department of Defense (DoD)’s Joint Logistics Enterprise, which
spans both supply chain and logistics operations, provides the means to muster, trans-
port, and sustain military power anywhere in the world at a high level of readiness.
To operate successfully in an increasingly contested global security environment, how-
ever, the logistics enterprise needs to change how it operates. As noted by the Defense
Science Board Task Force on Survivable Logistics, there is a critical need to “pro-
tect and enhance logistics information”, which is dispersed in thousands of disparate
legacy information systems, and provide inadequate visibility into the status of flow
of millions of military parts, supplies, and pieces of equipment, which are stocked and
shipped around the world. To address this challenge, the LogX program’s goal is to
develop and demonstrate software for real-time logistics and supply chain system sit-
uational awareness (diagnosis), future state prediction (prognosis), and assessment of
resilience at unprecedented scale and speed. LogX aims to build a capability to work
alongside existing logistics information systems that exploits the recent migration of
logistics information to digital formats and the cloud.

49
Materials Architectures and Char-
acterization for Hypersonics (MACH)
Dr. William Carter

https://www.darpa.mil/program/materials-architectures-and-characterization-for-hypersonics.html

GPT-3 Summary: GOAL: Develop new materials for sharp, shape-stable, cooled
leading edges for hypersonic vehicles.
TECHNOLOGY METHODS: Develop new materials using modern computation ca-
pabilities. Develop new passive and active thermal management concepts, coatings
and materials.
IMPACT: Enable new cooled hypersonic leading edge applications.

Full Details: The Materials Architectures and Characterization for Hypersonics

(MACH) program aims to develop and demonstrate new materials architectures for
sharp, shape-stable, cooled leading edges for hypersonic vehicles. The program will
investigate innovative approaches that enable revolutionary advances in the materi-
als, design and implementation of shape-stable, high heat flux capable leading edge
systems. MACH comprises two technical areas. The first area aims to develop and
mature fully integrated passive thermal management system to cool leading edges
based on scalable net-shape manufacturing and advanced thermal design. The sec-
ond technical area will focus on next-generation hypersonic materials research, apply-
ing modern high-fidelity computation capabilities to develop new passive and active
thermal management concepts, coatings and materials for future cooled hypersonic
leading edge applications.

50
Mission-Integrated Network Control
(MINC)
Dr. Mary R. Schurgot

https://www.darpa.mil/program/mission-integrated-network-control.html

GPT-3 Summary: GOAL: The goal of the MINC program is to ensure that critical
data can be routed to the correct user in highly contested and dynamic communica-
tion environments.
METHODS: The MINC program will develop an ”always-on” network overlay to ac-
cess available networking and communications resources and control parameters. It
will also use a cross-network approach for optimizing and managing network config-
uration and information flows.
IMPACT: This program has the potential to revolutionize how data is routed in
complex and contested environments.

Full Details: The objective of the MINC program is to ensure that critical data
finds a path to the right user at the right time in highly contested, highly dynamic
communication environments using secure control of any available communication or
networking resources (communications, compute, or storage capabilities). The MINC
approach is designed to interoperate with a heterogeneous mix of legacy and future
systems to ensure timely and reliable delivery of data that is not guaranteed today.
This capability replaces the manual, static configuration of individual, tactical net-
works and the associated limited internetworking capabilities. The growing emphasis
on all-domain warfare compounds the complexity of controlling heterogeneous net-
works. MINC embraces these challenges by introducing the mosaic warfare concepts
of optionality, diversity, and rapid adaptability to the orchestration of networks of
networks.The MINC program will realize ”on-demand” connectivity to enable effects
chains by focusing on the development and integration of three key capabilities: Cre-
ating an ”always-on” network overlay to access available networking and communica-
tions resources and control parameters;Using a cross-network approach for optimizing
and managing network configuration and information flows; andCreating a mission-
driven approach in determining the critical information flows for warfighting services.

51
Quantum Apertures (QA)
Dr. Jonathan Hoffman

https://www.darpa.mil/program/quantum-apertures.html

GPT-3 Summary: The Quantum Apertures (QA) program seeks to develop a new
way to receive radio-frequency (RF) signals using highly excited ”Rydberg” quan-
tum states. This will improve sensitivity and frequency agility for applications like
electronic warfare, radar, and communications. The QA program will exploit a phe-
nomenon called electromagnetically-induced transparency (EIT) to detect changes in
electron orbits caused by RF fields. This will allow for the development of portable
and directional RF receivers with greater sensitivity, bandwidth, and dynamic range.

Full Details: The Quantum Apertures (QA) program aims to develop a funda-
mentally new way of receiving radio-frequency (RF) waveforms to improve both
sensitivity and frequency agility in several application areas of interest to national
security including electronic warfare, radar, and communications.First demonstrated
in the DARPA QuASAR program, the target quantum apertures receivers will sense
electric fields using highly excited, so-called “Rydberg” quantum states with high
quantum number (n) of approximately 100. High-n states have electrons that orbit
˜10,000 times further away from the proton than a ground-state atom, making them
highly sensitive to electric fields; effectively acting like small antennae.In order to
detect the electric-field-induced changes in electron orbits caused by the RF electric
fields, the QA program seeks to exploit a phenomenon called electromagnetically-
induced transparency (EIT) that changes the intensity of a laser passing through a
cloud of Rydberg atoms. The QA program aims to develop portable and directional
RF receivers useful for future DoD missions with greater sensitivity, bandwidth, and
dynamic range than any classical receiver.

52
Hyper-wideband Enabled RF Mes-
saging (HERMES)
Dr. Tom Rondeau

https://www.darpa.mil/program/hyper-wideband-enabled-rf-messaging.html

GPT-3 Summary: GOAL: The HERMES program will develop a jammer-resistant

RF link that is orders of magnitude beyond the state-of-the-art.

TECHNOLOGY METHODS: The program will focus on systems that work

with extremely wide RF bandwidths, using processing gain, integrated filters, and
active cancellation to thwart jamming signals.

IMPACT: The HERMES program could provide unprecedentedly reliable channels

of access to the RF spectrum without pre-planned spectral allocation, allowing for a
new model of communications.

Full Details: Modern networks and platforms rely on access to the radio frequency
(RF) spectrum for communications, radar sensing, command and control, time trans-
fer, and geo-location. Electromagnetic interference, due to congestion in the spectrum
or malicious jamming, can have catastrophic effects. Countering such interference is
particularly important for unmanned platforms. To address this challenge, the Hyper-
wideband Enabled RF Messaging (HERMES) program seeks to provide an assured
link for essential communications by developing a jammer-countering capability that
is orders of magnitude beyond the state-of-the-art. To achieve this dramatic leap
in jam-resistance, the HERMES program is focusing on systems that work with ex-
tremely wide RF bandwidths. The goal of the program is to implement a robust
wideband (>10 GHz instantaneous bandwidth) RF link residing below 20GHz. The
link will operate at a 100 kb/s rate while deploying several jammer rejection tech-
niques—involving processing gain, integrated filters, and active cancellation—that
combine to thwart jamming signals that are even 10 million times stronger than the
communications signal. The broad spectral spreading of the signal will assure that
the transmit power in any spectral bin will be insufficient to cause unintentional
interference. If successful, the HERMES program could provide unprecedentedly re-
liable channels of access to the RF spectrum without pre-planned spectral allocation,
allowing for a new model of communications.

53
Software Defined Radio (SDR) 4.0
Dr. Tom Rondeau

https://www.darpa.mil/program/software-defined-radio.html

GPT-3 Summary: The ”Software Defined Radio (SDR) 4.0” program by DARPA
seeks to improve signal processing capabilities for SDRs through optimizing how the
GNU radio scheduler controls allocation of memory buffers and by improving signal
processing capabilities available in GNU Radio. This will enable faster and more
power-efficient computation of the specific signal-related mathematics and therefore
lower the overall power consumption of the system. This will have impacts in both
the government and commercial world, enabling new innovations with broad customer
appeal.

Full Details: SDR and software development kits (SDK) such as GNU Radio ex-
ist as free and open source technologies that are widely used in research, industry,
academia, government, and hobbyist environments to support both wireless com-
munications research and real-world radio systems. However, even with high end
multi-core x86 central processing units (CPU) there are adaptive radar, electronic
warfare (EW), and communications applications that cannot be implemented onto
SDR with a purely homogeneous CPU due to high latency and power consumption.
Furthermore coprocessors like field programmable gate arrays (FPGA) and graphics
processing units (GPU) are difficult to program, and prior attempts abstract them
to support direct use of open source SDKs have failed to achieve sufficient processing
efficiency to support more complex waveforms and protocols. Therefore, SDR 4.0
will develop technology that significantly improves the ability for SDRs to process
radio frequency signals and efficiently execute flexible, easy-to-program open source
software. The research effort will optimize how the GNU radio scheduler controls
allocation of memory buffers in order to improve the efficiency of data transfer to
and from coprocessor devices. Improving data transfer efficiency will significantly
improve signal processing capabilities and enhance mission performance associated
with radio frequency applications. The research effort will also improve signal pro-
cessing capabilities available in GNU Radio for use in heterogeneous processing stacks
that include a mix of CPUs, FPGAs, and GPUs. Offloading key signal processing
tasks to an FPGA or GPU while managing memory input/output will enable faster
and more power-efficient computation of the specific signal-related mathematics and
therefore lower the overall power consumption of the system. SDR 4.0 will develop an
improved foundation of SDR processing for use in both government and commercial
research. It is not creating new hardware or developing the means of expanding SDR

54
multi-functionality. In the commercial world, accelerated processing will enable GNU
Radio open–source software users to innovate on a whole host of new ideas with broad
customer appeal. Commercial research coupled with DoD research positively impacts
numerous defense applications of SDRs like those mentioned previously. For DARPA,
diverse radio functionality enabled by improved signals processing contributes to ad-
vanced Mosaic Warfare objectives such as system recomposability.

55
Synergistic Discovery and Design (SD2)
Dr. Joshua Elliott

https://www.darpa.mil/program/synergistic-discovery-and-design.html

GPT-3 Summary: GOAL: Develop data-driven methods to accelerate scientific

discovery and robust design in domains that lack complete models.

TECHNOLOGY METHODS: SD2 will address the problem of design in do-

mains that lack complete models by discovering models and refining designs via
methods that extract information from data at petabyte scale.

IMPACT: Provide new data-driven methods to accelerate discovery and design

and create a cloud-based open data exchange for research communities in complex
domains.

Full Details: The Synergistic Discovery and Design (SD2) program aims to develop
data-driven methods to accelerate scientific discovery and robust design in domains
that lack complete models. Engineers regularly use high-fidelity simulations to create
robust designs in complex domains such as aeronautics, automobiles, and integrated
circuits. In contrast, robust design remains elusive in domains such as synthetic
biology, neuro-computation, and polymer chemistry due to the lack of high-fidelity
models. SD2 seeks to develop tools to enable robust design despite the lack of com-
plete scientific models. Examples of complex systems where inventors lack complete
scientific models to support their design efforts include biological systems that have
millions of protein-metabolite interactions, neuro-processes that require computa-
tions across billions of neurons, and advanced materials influenced by millions of
monomer-protein combinations. These systems are part of domains that exhibit mil-
lions of unpredictable, interacting components for which robust models do not exist,
and internal states are often only partially observable. In such domains, small pertur-
bations in the environment can lead to unexpected design failures, and the number
of engineering variables required to characterize stable operational envelopes remains
unknown. While domain experts remain geographically dispersed, they collectively
analyze hundreds of terabytes of data to build models and refine designs. However,
manually-intensive analysis of small datasets remains inefficient and often yields unre-
producible results. In response, researchers have begun to outsource high-throughput
experiments to automated labs and randomly search constrained parameter spaces for
robust designs. However, these random search-based approaches work best with small
parameter spaces and provide no insight into why some designs succeed and others

56
fail. SD2 will address the problem of design in domains that lack complete models by
discovering models and refining designs via methods that extract information from
data at petabyte scale. SD2 aims to develop data-driven methods to automatically
discover models and refine designs in parallel at scale. To ensure realism, challenge
problems drawn from cutting edge domains will drive the development of SD2 meth-
ods. By the end of the program, SD2 plans to provide new data-driven methods
to accelerate discovery and design and create a cloud-based open data exchange for
research communities in complex domains.

57
In the Moment (ITM)
Dr. Matt Turek

https://www.darpa.mil/program/in-the-moment.html

GPT-3 Summary: GOAL: Develop algorithms that can be trusted to make deci-
sions independently in difficult domains, such as military operations.
TECHNOLOGY METHODS: Identify key attributes of trusted human decision-
making and represent them computationally. Generate a quantitative alignment
framework for a trusted human decision-maker and an algorithm.
IMPACT: Enable algorithms to be tuned to align with specific, trusted humans, and
develop autonomous decision-making systems.

Full Details: In the Moment (ITM) aims to support the development of algorithms
that are trusted to independently make decisions in difficult domains. Military oper-
ations – such as mass casualty triage and disaster relief – require complex and rapid
decision-making in dynamic situations where there is often no human consensus and
no ground truth. By identifying key attributes underlying trusted human decision-
making in dynamic settings and computationally representing those attributes, the
ITM program seeks to generate a quantitative alignment framework for a trusted
human decision-maker and an algorithm. This alignment framework will form the
basis for understanding algorithm alignment with trusted humans, enable algorithms
to incorporate key trusted attributes, and support the development of algorithms
that can be tuned to align with specific, trusted humans. A dedicated team will
provide ethical, legal, and social issues guidance, as well as consider future policy
options for autonomous decision-making systems.ITM will investigate these decision-
making problems in the context of two specific domains: small unit triage in austere
environments in Phase 1 and mass casualty triage in Phase 2. The ITM program
will encourage cross-disciplinary research in the fields of decision-making, cognitive
science, experimental psychology, simulation environments, artificial intelligence, ma-
chine learning, algorithm evaluation, and decision-making policy for autonomous sys-
tems.HR001122S0031: In the Moment (ITM)

58
Neural Engineering System Design
(NESD)
Dr. Joeanna Arthur

https://www.darpa.mil/program/neural-engineering-system-design.html

GPT-3 Summary: The Neural Engineering System Design (NESD) program seeks
to develop high-resolution neurotechnology to improve the function of the visual and
auditory systems for military personnel who have been injured. The program uses
a combination of hardware and software to create an interface that can read and
write to a large number of neurons. If successful, the program has the potential to
significantly improve treatments for sensory deficits and to provide new understanding
of the neural underpinnings of vision, hearing, and speech.

Full Details: The Neural Engineering System Design (NESD) program seeks to de-
velop high-resolution neurotechnology capable of mitigating the effects of injury and
disease on the visual and auditory systems of military personnel. In addition to creat-
ing novel hardware and algorithms, the program conducts research to understand how
various forms of neural sensing and actuation might improve restorative therapeutic
outcomes. The focus of the program is development of advanced neural interfaces that
provide high signal resolution, speed, and volume data transfer between the brain and
electronics, serving as a translator for the electrochemical language used by neurons
in the brain and the ones and zeros that constitute the language of information tech-
nology. The program aims to develop an interface that can read 106 neurons, write
to 105 neurons, and interact with 103 neurons full-duplex, a far greater scale than is
possible with existing neurotechnology.To succeed, NESD requires integrated break-
throughs across disciplines including neuroscience, low-power electronics, photonics,
medical device packaging and manufacturing, systems engineering, and clinical test-
ing. In addition to hardware, NESD performer teams are developing advanced math-
ematical and neuro-computation techniques to first transcode high-definition sensory
information between electronic and cortical neuron representations and then com-
press and represent those data with minimal loss of fidelity and functionality. If the
program is successful, the work has the potential to significantly advance scientists’
understanding of the neural underpinnings of vision, hearing, and speech and could
eventually lead to new treatments for injured Service members living with sensory
deficits. Additionally, NESD tools could yield new understanding of the architecture
and processing of highly integrated neural circuits.

59
Assured Autonomy
Mr. William Martin

https://www.darpa.mil/program/assured-autonomy.html

GPT-3 Summary: The Assured Autonomy program will focus on creating technol-
ogy for the continual assurance of Learning-Enabled, Cyber Physical Systems (LE-
CPSs). This will be done by developing tools and toolchains that can be integrated
into LE-CPS design processes. The goal is to provide greater assurance of safety and
correctness for these systems as they evolve over time.

Full Details: Autonomy refers to a system’s ability to accomplish goals indepen-

dently, or with minimal supervision from human operators in environments that are
complex and unpredictable. Autonomous systems are increasingly critical to several
current and future Department of Defense (DoD) mission needs. For example, the
U.S. Army Robotics and Autonomous Systems (RAS) strategy report for 2015-2040
identifies a range of capability objectives, including enhanced situational awareness,
cognitive workload reduction, force protection, cyber defense, logistics, etc, that rely
on autonomous systems and higher levels of autonomy. Tremendous advances have
been made in the last decade in constructing autonomous Cyber Physical Systems
(CPS), as evidenced by the proliferation of a variety of unmanned systems: air,
ground, sea, and undersea vehicles. These advances have been driven by innovations in
several areas, such as sensor and actuator technologies, computing technologies, con-
trol theory, design methods and tools, modeling and simulation technologies, among
others. In spite of these advances, deployment and broader adoption of such systems
in safety-critical DoD applications remains challenging and controversial.Several fac-
tors impede the deployment and adoption of autonomous systems: In the absence of
an adequately high level of autonomy that can be relied upon, substantial operator
involvement is required, which not only severely limits operational gains, but creates
significant new challenges in the areas of human-machine interaction and mixed ini-
tiative control. Achieving higher levels of autonomy in uncertain, unstructured, and
dynamic environments, on the other hand, increasingly involves data-driven machine
learning techniques with many open systems science and systems engineering chal-
lenges.Machine learning techniques widely used today are inherently unpredictable
and lack the necessary mathematical framework to provide guarantees on correct-
ness, while DoD applications that depend on safe and correct operation for mission
success require predictable behavior and strong assurance.Historically, assurance has
been approached through design processes following rigorous safety standards in de-
velopment, and demonstrating compliance through system testing. However, these

60
standards have been developed primarily for human-in-the-loop systems, and are
bounded in scope, not extending to advanced levels of autonomy where system be-
havior depends on its memory of received stimuli. Current assurance approaches are
predicated on the assumption that once the system is deployed, it does not learn and
evolve. The goal of the Assured Autonomy program is to create technology for contin-
ual assurance of Learning-Enabled, Cyber Physical Systems (LE-CPSs). Continual
assurance is defined as an assurance of the safety and functional correctness of the
system provided provisionally at design time, and continually monitored, updated,
and evaluated at operation-time as the system and its environment evolves. An LE-
CPS is defined as a system composed of one or more Learning-enabled Components
(LECs). A LEC is a component whose behavior is driven by “background knowledge”
acquired and updated through a “learning process,” while operating in a dynamic and
unstructured environment. This definition generalizes and admits a variety of popular
machine learning approaches and algorithms (e.g., supervisory learning for training
classifiers, reinforcement learning for developing control policies, algorithms for learn-
ing system dynamics). The generalization is intentional to promote abstractions and
tools that can be applied to different types and applications of data-driven machine
learning algorithms in Cyber Physical Systems (CPSs) to enhance their autonomy. In
order to ground the Assured Autonomy research objectives, the program will priori-
tize challenge problems in the militarily relevant autonomous vehicle space. However,
it is anticipated that the tools, toolchains, and algorithms created will be relevant to
other LE-CPSs. The resulting technology from the program will be in the form of a
set of publicly available tools integrated into LE-CPS design toolchains that will be
made widely available for use in commercial and defense sectors.

61
Gambit
Maj. Nathan Greiner, USAF

https://www.darpa.mil/program/gambit.html

GPT-3 Summary: GOAL: The Gambit program focuses on developing Rotating

Detonation Engines (RDEs) as a new class of propulsion to enable standoff strike of
time-critical targets from fourth generation fighters at campaign scale.
METHODS: Phase 1 focuses on preliminary design of the freejet test article and sub-
stantiating the design with combustor and inlet testing. Phase 2 focuses on detailed
design, fabrication, and testing of the freejet test article.
IMPACT: Gambit could lay the foundation for future potential prototype weapon
development.

Full Details: The Gambit program is focused on developing Rotating Detonation

Engines (RDEs) as a new class of propulsion to enable standoff strike of time-critical
targets from fourth generation fighters at campaign scale. Current propulsion tech-
nologies (e.g. conventional rockets, ramjets, and gas-turbine engines) have limitations
in maximum range, speed, and/or affordability that hinders their ability to meet the
Gambit program objectives. RDEs are more compact than conventional ramjets,
which allows them to carry more fuel. Additionally, RDEs are less complex than
gas turbine engines. The combination of these two factors enable RDEs to afford-
ably meet the Gambit program objectives.Gambit will develop and demonstrate RDE
technology in a full-scale freejet test. Phase 1 focuses on preliminary design of the
freejet test article and substantiating the design with combustor and inlet testing.
Phase 2 focuses on detailed design, fabrication, and testing of the freejet test article.
Gambit could lay the foundation for future potential prototype weapon development.

62
Electrical Prescriptions (ElectRx)
Dr. Gopal Sarma

https://www.darpa.mil/program/electrical-prescriptions.html

GPT-3 Summary: GOAL: The goal of the Electrical Prescriptions (ElectRx) pro-
gram is to reduce the time to treatment, logistical challenges, and potential off-target
effects associated with traditional medical interventions for a wide range of physical
and mental health conditions commonly faced by our warfighters.
METHODS: This program seeks to deliver non-pharmacological treatments for pain,
general inflammation, post-traumatic stress, severe anxiety, and trauma that em-
ploy precise, closed-loop, non-invasive modulation of the patient’s peripheral nervous
system in order to exploit and supplement the body’s natural ability to quickly

Full Details: The Electrical Prescriptions (ElectRx) program aims to support mil-
itary operational readiness by reducing the time to treatment, logistical challenges,
and potential off-target effects associated with traditional medical interventions for a
wide range of physical and mental health conditions commonly faced by our warfight-
ers. ElectRx seeks to deliver non-pharmacological treatments for pain, general in-
flammation, post-traumatic stress, severe anxiety, and trauma that employ precise,
closed-loop, non-invasive modulation of the patient’s peripheral nervous system. The
human nervous system already plays a vital role in maintaining all aspects of physical
and mental health. A sophisticated network of sensory nerves continuously monitors
health status and triggers reflexive responses in the brain and spinal cord when an in-
fection or injury is detected. These reflexes normally adjust organ function to initiate
and control the healing process. However, some diseases can disrupt healthy func-
tioning of these processes and produce nerve signaling that causes pain, metabolic
disorders such as diabetes, and autoimmune disorders such as rheumatoid arthri-
tis. ElectRx technology would exploit and supplement the body’s natural ability
to quickly and effectively heal itself, intervening when required to correct or bolster
nervous system activity. ElectRx is establishing the underlying science and develop-
ing the technologies that could enable artificial modulation of peripheral nerves to
restore healthy patterns of signaling in these neural circuits. The program seeks to
advance understanding of the anatomy and physiology of specific neural circuits and
their role in health and disease. In parallel, the program also seeks to develop novel
biological-interface technologies for monitoring biomarkers and peripheral nerve ac-
tivity, and delivering therapeutic signals to peripheral nerve targets. Potential new
approaches include in vivo, real-time biosensors and novel neural interfaces using
optical, acoustic, electromagnetic, or engineered biology strategies to achieve precise

63
targeting with potentially single-axon resolution.Following successes in early proof-of-
concept studies, the ElectRx devices and therapeutic systems under development are
entering into clinical studies. If successful, such precise neuromodulation capability
technology would reduce dependence on traditional drugs and create new treatments
that could be automatically and continuously tuned to the needs of warfighters with-
out side effects. The technology could also help doctors evaluate and predict various
physiological states, and characterize host response in patients with severe infections,
providing a quantitative framework to guide operations and therapy.

64
Causal Exploration of Complex Op-
erational Environments (Causal Ex-
ploration)
Dr. Joshua Elliott

https://www.darpa.mil/program/causal-exploration.html

GPT-3 Summary: GOAL: Develop a tool to help military planners understand

and address the underlying causal factors that drive complex conflict situations.
TECHNOLOGY METHODS: Develop a modeling platform that enables users to
rapidly create, maintain, and interact with a causal model tailored for the operational
environment they are facing.
IMPACT: Help users gain in-depth understanding of the operational environment to
support and inform their planning efforts.

Full Details: Over the last 15 years, the U.S. military has increasingly been called
upon to face complex operational environments (OE) and diverse enemies. Such com-
plex OEs require the actions of U.S. forces and host-nation or coalition partners to be
based on a common holistic understanding of the OE (e.g., governments, population
groups, security forces, and violent non-state actors) and, in particular, the causal
dynamics that can manifest as unanticipated and often counter-intuitive outcomes.
While a range of modeling and simulation tools exist within military commands, most
are special-purpose and extensive time and effort are often required to configure and
use these tools. Existing tools are also heavily dependent on databases with no au-
tomated mechanism for keeping them current. These tools are not generally suitable
for use directly by operational planners as they require expert modelers to assemble,
configure, run, and interpret the outputs. Causal Exploration seeks to develop a mod-
eling platform to aid military planners in understanding and addressing underlying
causal factors that drive complex conflict situations. The technologies embodied in
the Causal Exploration platform will enable users to rapidly create, maintain, and
interact with a causal model that has been tailored for the operational environment
they are facing. Interaction with the model will allow users to explore the causal
dynamics driving the conflict, and gain in-depth understanding of the operational
environment to support and inform their planning efforts. While this capability will
have broad applicability, the program will focus on hybrid or irregular conflicts, which
are dominated by complex human dynamics with intertwining political, territorial,
economic, ethnic, and/or religious tensions.

65
Environment-driven Conceptual Learn-
ing (ECOLE)
Dr. William Corvey

https://www.darpa.mil/program/environment-driven-conceptual-learning.html

GPT-3 Summary: GOAL: The ECOLE program will develop artificial intelligence
agents that can learn from language and vision to enable human-machine collabora-
tive analysis of images, videos, and multimedia documents.

METHODS:Develop algorithms that can identify, represent, and ground novel

attributes that form the symbolic and contextual model for a particular object or
activity through interactive learning with a human analyst.

IMPACT:The ECOLE program will transform current ML approaches by de-

veloping algorithms that can identify, represent, and ground novel attributes that
form the symbolic and contextual model for a particular object or activity through
interactive learning with a human analyst.

Full Details: The United States Department of Defense (DoD) and Intelligence
Community (IC) need computational systems that can robustly and automatically
analyze large amounts of multimodal data. Furthermore, these computational sys-
tems need to be able to communicate and cooperate with human beings to resolve
ambiguities and improve performance over time. The Environment-driven Conceptual
Learning (ECOLE) program will create artificial intelligence (AI) agents capable of
continually learning from language and vision to enable human-machine collaborative
analysis of image, video, and multimedia documents during time-sensitive, mission-
critical DOD analytic tasks, where reliability and robustness are essential.ECOLE
will transform current ML approaches by developing algorithms that can identify,
represent, and ground novel attributes that form the symbolic and contextual model
for a particular object or activity through interactive learning with a human ana-
lyst.ECOLE is a 48-month, three-phase program with 18-month Phase 1 and Phase
2 efforts and a 12-month, transition-focused Phase 3 concentrated specifically on de-
velopment concerns related to GEOINT workflows.

66
Next-Generation Nonsurgical Neu-
rotechnology
Dr. Gopal Sarma

https://www.darpa.mil/program/next-generation-nonsurgical-neurotechnology.html

GPT-3 Summary: GOAL: Develop a new neurotechnology that does not require
surgery, is man-portable, and can be used by a wider population.

TECHNOLOGY METHODS: Develop an integrated device that can read from and
write to 16 independent channels within a 16mm3 volume of neural tissue.

IMPACT: Enable non-invasive brain-machine interfaces that have the precision

to rival existing invasive approaches.

Full Details: The Next-Generation Nonsurgical Neurotechnology (N3) program

aims to develop high-performance, bi-directional brain-machine interfaces for able-
bodied service members. Such interfaces would be enabling technology for diverse
national security applications such as control of unmanned aerial vehicles and active
cyber defense systems or teaming with computer systems to successfully multitask
during complex military missions. Whereas the most effective, state-of-the-art neural
interfaces require surgery to implant electrodes into the brain, N3 technology would
not require surgery and would be man-portable, thus making the technology accessi-
ble to a far wider population of potential users. Noninvasive neurotechnologies such
as the electroencephalogram and transcranial direct current stimulation already exist,
but do not offer the precision, signal resolution, and portability required for advanced
applications by people working in real-world settings.The envisioned N3 technology
breaks through the limitations of existing technology by delivering an integrated de-
vice that does not require surgical implantation, but has the precision to read from
and write to 16 independent channels within a 16mm3 volume of neural tissue within
50ms. Each channel is capable of specifically interacting with sub-millimeter regions
of the brain with a spatial and temporal specificity that rivals existing invasive ap-
proaches. Individual devices can be combined to provide the ability to interface to
multiple points in the brain at once. To enable future non-invasive brain-machine
interfaces, N3 researchers are working to develop solutions that address challenges
such as the physics of scattering and weakening of signals as they pass through skin,
skull, and brain tissue, as well as designing algorithms for decoding and encoding
neural signals that are represented by other modalities such as light, acoustic, or

67
electro-magnetic energy.

68
Sea Train
Dr. Andrew Nuss

https://www.darpa.mil/program/sea-train.html

GPT-3 Summary: The Sea Train program seeks to develop and demonstrate ap-
proaches for increasing the range of medium unmanned surface vessels by physically
connecting vessels or sailing in collaborative formations. If successful, this could
enable extended transoceanic transit and long-range naval operations.

Full Details: The Sea Train program aims to demonstrate long range deployment
capabilities for a distributed fleet of tactical unmanned surface vessels. The program
seeks to enable extended transoceanic transit and long-range naval operations by
exploiting the efficiencies of a system of connected vessels (Sea Train). The goal is to
develop and demonstrate approaches that exploit wave-making resistance reductions
to overcome the range limitations inherent in medium unmanned surface vessels.
DARPA envisions sea trains formed by physically connecting vessels with various
degrees of freedom between the vessels, or vessels sailing in collaborative formations
at various distances between the vessels.

69
Photonics in the Package for Extreme
Scalability (PIPES)
Dr. Gordon Keeler

https://www.darpa.mil/program/photonics-in-the-package-for-extreme-scalability.html

GPT-3 Summary: GOAL: The Photonics in the Package for Extreme Scalability
(PIPES) program will use light to vastly improve the digital connectivity of packaged
integrated circuits.
TECHNOLOGY METHODS: The program will integrate state-of-the-art electronics
with photonic modules that deliver 100 Tbps per package.
IMPACT: The technologies developed through this program will have a major impact
on a host of dual-use applications, including artificial intelligence, machine learning,
large scale emulation, and high performance computing.

Full Details: Advances in digital microelectronics have enabled indispensable capa-

bilities for the Department of Defense (DoD) in the fields of information processing,
sensors, and communications. Increasingly, system performance in these domains
is constrained not by the limits of computation at individual nodes, but by elec-
trical data movement between individual chips. The Photonics in the Package for
Extreme Scalability (PIPES) program is using light to vastly improve the digital
connectivity of packaged integrated circuits, with a goal of enabling revolutionary
performance gains through parallelism and modularity. PIPES seeks to embed op-
tical signaling technologies within the package of application-specific integrated cir-
cuits (ASICs) and field-programmable gate arrays (FPGAs) to enable data links with
unprecedented bandwidth density, efficiency, and reach. Specifically, PIPES aims to
integrate state-of-the-art electronics with photonic modules that deliver 100 Tbps per
package at energies less than 1 picojoule per bit, while creating technologies capable
of 10x higher performance for future applications. As PIPES technologies mature,
DARPA anticipates pervasive adoption of embedded optics for signaling in advanced
microelectronic systems. Beyond FPGAs and specialty DoD products, photonic con-
nectivity for central processing units (CPUs), graphical processing units (GPUs), and
domain-specific accelerators will impact a host of dual-use applications, including ar-
tificial intelligence, machine learning, large scale emulation, and high performance
computing.

70
Squad X
Dr. Stuart H. Young

https://www.darpa.mil/program/squad-x.html

GPT-3 Summary: GOAL: The Squad X program will design, develop, and vali-
date autonomous system prototypes to help ensure U.S. small unit dominance over
adversaries in the future.
TECHNOLOGY METHODS: Develop technologies in four key areas: precision en-
gagement, non-kinetic engagement, squad sensing, and squad autonomy.
IMPACT: Improve situational understanding of the multi-domain environment; op-
timize squad maneuverability; and shape and dominate the battlespace.

Full Details: Modern military engagements increasingly take place in complex and
uncertain battlefield conditions where attacks can come from multiple directions at
once, and in the electromagnetic spectrum and cyber domains, as well. U.S. Army
and U.S. Marine Corps dismounted infantry squads have been unable to take full
advantage of some highly effective multi-domain defensive and offensive capabilities
that vehicle-assigned forces currently enjoy – in large part because many of the rele-
vant technologies are too heavy and cumbersome for individual warfighters to carry
or too difficult to use under demanding field conditions. DARPA’s Squad X seeks
to help overcome these challenges and help ensure U.S. small unit dominance over
adversaries in the decades to come. The goal of the Squad X Experimentation pro-
gram is to design, develop, and validate autonomous system prototypes, equipped
with novel sensing tools and off-the-shelf technologies, some of which were identi-
fied via the previous Squad X Core Technologies program. The technologies aim to
increase squads’ situational awareness, while the autonomous systems allow squads
to increase their battle space and area of influence. Squad X focuses its technol-
ogy development in four key areas: Precision Engagement: Precisely engage threats
while maintaining compatibility with infantry weapon systems and without impos-
ing weight or operational burdens that would negatively affect mission effectiveness.
Capabilities of interest include distributed, non-line-of-sight targeting and guided
munitions. Non-Kinetic Engagement: Disrupt enemy command and control, commu-
nications and use of unmanned assets at a squad-relevant operational pace (walking
with occasional bursts of speed). Capabilities of interest include disaggregated elec-
tronic surveillance and coordinated effects from distributed platforms.Squad Sensing:
Detect potential threats at a squad-relevant operational pace. Capabilities of interest
include multi-source data fusion and autonomous threat detection.Squad Autonomy:
Increase squad members’ real-time knowledge of their own and teammates’ locations

71
in GPS-denied environments through collaboration with embedded unmanned air and
ground systems. Capabilities of interest include robust collaboration between humans
and unmanned systems.The resulting integrated systems would improve situational
understanding of the multi-domain environment; optimize squad maneuverability via
physical, cognitive and material resources; and shape and dominate the battlespace
via synchronization of fire and maneuvering in all three domains: physical, electro-
magnetic, and cyber.

72
Measuring Biological Aptitude (MBA)
Dr. Gopal Sarma

https://www.darpa.mil/program/measuring-biological-aptitude.html

GPT-3 Summary: The MBA program’s goal is to help warfighters improve their
performance by measuring and tracking personalized biomarkers related to train-
ing and peak performance. The program’s technology methods include developing
new devices to track expression circuits related to cognitive, behavioral, and physi-
cal traits. If successful, the MBA program would provide a complete picture of the
warfighter, tying performance traits to signaling networks in the body. This would en-
able both individual warfighters and the Department of Defense to better understand
what to measure during training, mission execution, and recovery.

Full Details: Whereas the tools and weapons that are used by our warfighters have
evolved dramatically in the past few decades, the way in which the warfighter is pre-
pared has not kept pace with those developments. The Measuring Biological Aptitude
(MBA) program aims to address the need for a more capable fighting force by helping
individual warfighters identify, measure, and track personalized biomarkers related to
training and peak performance for specialized roles. If the program succeeds, MBA
technologies will give warfighters the ability to understand the underlying biological
processes that govern their performance. Specifically, these technologies would elu-
cidate the internal expression circuits (e.g., genetic, epigenetic, metabolomic) that
shape militarily relevant cognitive, behavioral, and physical traits. New devices for
continuously tracking these expression circuits could be integrated into the body to
provide instantaneous user feedback, helping the warfighter to improve performance
throughout training, assessment, selection, and mission execution for a given mil-
itary specialty.Successful MBA technologies would change how the Department of
Defense understands individual warfighters, updating a more than 50-year-old evalu-
ation system that is based on inadequate screening and aptitude tests supplemented
by annual medical examinations. This long-awaited modernization of performance
measurement would provide a complete picture of the warfighter, tying performance
traits to myriad signaling networks in the body that drive specific aspects of perfor-
mance. MBA could ultimately enable both individual warfighters and the department
to fully understand what to measure during training, mission execution, and recovery
to maximize long-term success in specialized roles.

73
Thermal Engineering using Materi-
als Physics (TEMP)
Dr. William Carter

https://www.darpa.mil/program/thermal-engineering-using-metamaterial.html

GPT-3 Summary: GOAL: The Thermal Engineering using Metamaterial Physics

(TEMP) program will develop materials and physics to control the direction and
wavelength of thermal transport by radiation in the visible spectrum in extreme ther-
mal conditions.
TECHNOLOGY METHODS: Develop solutions that can adaptively control the
directionality of radiative heat transfer while simultaneously withstanding high-
temperature oxidative or corrosive environments.
IMPACT: The program will demonstrate that proposed solutions are manufacturable
and scalable to be commercially relevant.

Full Details: The Thermal Engineering using Metamaterial Physics (TEMP) Dis-
ruption Opportunity (DO) leverages recent developments in materials and physics to
control the direction and wavelength of thermal transport by radiation in the visi-
ble spectrum in extreme thermal conditions. Solutions are being developed that can
adaptively control the directionality of radiative heat transfer while simultaneously
withstanding high-temperature oxidative or corrosive environments.Applications for
TEMP technology include thermophotovoltaic energy conversion and turbine engines.
The program will also demonstrate that proposed solutions are manufacturable and
scalable to be commercially relevant.

74
Hand Proprioception and Touch In-
terfaces (HAPTIX)
Dr. Matthew Pava

https://www.darpa.mil/program/hand-proprioception-and-touch-interfaces.html

GPT-3 Summary: The goal of the HAPTIX program is to develop a system that
would provide users of upper-limb prosthetic devices with near-natural control and
sensory feedback. The program is focusing on developing technologies to interface
permanently and continuously with the peripheral nerves in humans in order to tap
into the motor and sensory signals of the arm. If successful, the HAPTIX system
would be integrated with an advanced prosthetic limb to create a dexterous prosthetic
limb with full sensory and motor capabilities. The anticipated impact of this program
is the creation of a prosthetic limb that is suitable for home use and that can

Full Details: With a focus on wounded warriors and facilitating their return to mil-
itary service, the Hand Proprioception and Touch Interfaces (HAPTIX) program
is pursuing key technologies to enable precision control of and sensory feedback
from sensor-equipped upper-limb prosthetic devices. If successful, the resulting sys-
tem would provide users near-natural control of prosthetic hands and arms via bi-
directional peripheral nerve implants. The program has a strong focus on technology
handoff and aims to create and transition clinically relevant technology in support
of wounded warriors suffering from single or multiple limb loss. HAPTIX builds on
prior DARPA investments in the Reliable Neural-Interface Technology (RE-NET)
program, which created novel neural interface systems that overcame previous sensor
reliability issues to now last for the lifetime of the patient. A key focus of HAPTIX
is on creating new technologies to interface permanently and continuously with the
peripheral nerves in humans. HAPTIX technologies are being designed to tap into the
motor and sensory signals of the arm to allow users to control and sense the prosthesis
via the same neural signaling pathways used for intact limbs. Direct access to these
natural control signals will, if successful, enable more natural, intuitive control of com-
plex hand movements, and the addition of sensory feedback will further improve hand
functionality by enabling users to sense grip force and hand posture. Sensory feedback
may also provide important psychological benefits such as improving prosthesis “em-
bodiment” and reducing the phantom limb pain that is suffered by approximately 80
percent of amputees.In addition to developing the low-power microelectronics needed
for the system, HAPTIX performer teams also conduct fundamental neuroscience re-
search to understand how the nervous system encodes motor and sensory information

75
for the hand. This knowledge guides development of algorithms that enable intuitive
control of the prosthesis and provide rich sensations of touch and proprioception. If
successful, the completed HAPTIX system will be integrated with one of the advanced
prosthetic limbs developed under DARPA’s Revolutionizing Prosthetics program to
create the first dexterous prosthetic limb with full sensory and motor capabilities that
is suitable for home use. DARPA anticipates a 12-month, take-home clinical trial of
the complete HAPTIX system as the culmination of the program.

76
Dynamic Range-enhanced Electron-
ics and Materials (DREaM)
Dr. Thomas Kazior

https://www.darpa.mil/program/dynamic-range-enhanced-electronics-and-materials.html

GPT-3 Summary: GOAL: The DREaM program will create RF/mmW transistors
that enable asymmetric operations in a complex electromagnetic spectrum.

TECHNOLOGY METHODS: Develop non-traditional materials, integrate new

device structures, and innovate on transistor layouts.

IMPACT: Create high dynamic range RF transistors for a diverse set of am-
plifier applications.

Full Details: Over the past decade, DARPA’s investments in the advancement of
Gallium Nitride (GaN) technology have helped enable the delivery of high power radio
frequency (RF) signals at higher frequencies, bandwidths, and efficiencies. Today,
however, a growing number of commercial and military components – from everyday
smartphones to RF jammers – are generating a vast amount of RF signals, which is
creating an increasingly crowded electromagnetic environment and a need to utilize
higher operating frequencies – moving up to millimeter wave (mmW) frequencies.
To operate in this complex spectrum environment with large signal-to-noise ratios,
next-generation mmW RF systems will require high dynamic range. However, today’s
mmW transistor technologies are fundamentally limited by their ability to process RF
signals efficiently with large bandwidth and high fidelity at high operating powers.The
Dynamic Range-enhanced Electronics and Materials (DREaM) program is exploiting
new materials and novel device structures to create RF/mmW transistors that will
enable asymmetric operations in a complex electromagnetic spectrum. To achieve its
intended goals, DREaM is creating high dynamic range RF transistors for a diverse
set of amplifier applications by developing non-traditional materials, integrating new
device structures, and innovating on transistor layouts to attain a 2X higher output
power density and 10X better amplifier linearity than the state of the art devices of
today.

77
Robotic Autonomy in Complex En-
vironments with Resiliency (RACER)
Dr. Stuart H. Young

https://www.darpa.mil/program/robotic-autonomy-in-complex-environments-with-resiliency.html

GPT-3 Summary: GOAL: Develop new autonomy algorithm technologies for

Unmanned Ground Vehicles (UGVs) to maneuver on unstructured off-road terrain.

TECHNOLOGY METHODS: Develop software for autonomous on-road and

off-road missions. Implement them in UGVs. Test algorithms in the field.

IMPACT: Improve algorithms for UGVs to enable them to move off-road at

speeds on par with a human driver.

Full Details: The Robotic Autonomy in Complex Environments with Resiliency

(RACER) program is focused on developing and demonstrating new autonomy algo-
rithm technologies, rather than vehicle or sensor technologies, that enable Unmanned
Ground Vehicles (UGVs) to maneuver on unstructured off-road terrain at speeds that
are only limited by considerations of sensor performance, mechanical constraints, and
safety. At a minimum, the program goal is software performance to move off-road
at speeds on par with a human driver. The self-driving car industry has made rapid
advances via a vehicle platform-based, agile develop-test-develop-test model that has
accumulated data to help train algorithms and refine approaches. Simulation-based
development approaches use the same data for algorithm tuning. On-road auton-
omy algorithms operate in well-structured and highly predictable environments with
limited obstacles. However, military off-road autonomy algorithms and capability
development has lagged due to the challenging complexity of off-road terrain environ-
ments and need to travel in them at relevant speeds.While RACER seeks to leverage
advances in on-road autonomy, the project is investigating innovative approaches that
enable revolutionary progress in algorithms operating on systems rather than science
or research that primarily results in evolutionary improvements to the existing state
of practice. RACER will demonstrate game-changing autonomous UGV mobility,
focused on speed and resiliency, using a combination of simulation and advanced
platforms. It tests algorithms in the field at DARPA-hosted experiments across the
country on a variety of terrain. The program provides UGV platforms that research
teams can use to develop autonomous software capabilities through repeated cycles
of tests on unstructured off-road landscapes. Goals include not only autonomy algo-

78
rithms, but also creation of simulation-based approaches and environments that will
support rapid advancement of self-driving capabilities for future UGVs.

79
Novel Orbital and Moon Manufac-
turing, Materials, and Mass-efficient
Design (NOM4D)
Dr. William Carter

https://www.darpa.mil/program/novel-orbital-and-moon-manufacturing-materials-and-
mass-efficient-design.html

GPT-3 Summary: GOAL: Develop the ability to build structures in space or on

the moon, with a focus on solar arrays, antennas, and optics.
TECHNOLOGY METHODS: Develop and demonstrate foundational materials, pro-
cesses, and designs to enable on-orbit and on-moon fabrication of structures.
IMPACT: Investigate designs that take advantage of space-based manufacturing to
create mass-efficient structures that can withstand the space and lunar environment.

Full Details: The Novel Orbital and Moon Manufacturing, Materials, and Mass-
efficient Design (NOM4D, pronounced NOMAD) program aims to develop the foun-
dations of building robust, precise structures in space or on the moon. In contrast to
today’s deployable structures that are optimized for ground test and launch survival,
structures such as solar arrays, antennas or optics will be specifically designed for the
space or lunar environment. The NOM4D program comprises two technical areas.
The first plans to develop and demonstrate foundational materials, manufacturing
processes, and designs to enable the on-orbit and on-moon fabrication of robust, re-
silient, and high-precision structures that will support future off-earth space systems.
The second technical area will investigate innovative designs that take advantage of
the ability to manufacture in space, yet enable precise, mass-efficient future space
structures that withstand maneuvers, eclipses, damage, and thermal cycles inherent
to the space and lunar environments. The goal is to do so with mass efficiencies that
transcend the limits of today’s stiffness-driven designs.

80
Magnetic Miniaturized and Mono-
lithically Integrated Components (M3IC)
Dr. David K. Abe

https://www.darpa.mil/program/magnetic-miniaturized-and-monolithically-integrated-
components.html

GPT-3 Summary: GOAL: The M3IC program will develop technology to integrate
magnetic components onto semiconductor materials.
TECHNOLOGY METHODS: This will be done by improving the size and func-
tionality of electromagnetic (EM) systems for communications, radar, and electronic
warfare (EW).
IMPACT: The overall goal is to reduce the size, weight, and power (SWaP) of mag-
netic components and integrating them onto semiconductor chips.

Full Details: Magnetic Miniaturized and Monolithically Integrated Components

(M3IC) program aims to integrate magnetic components onto semiconductor mate-
rials, improving the size and functionality of electromagnetic (EM) systems for com-
munications, radar, and electronic warfare (EW). Current EM systems use magnetic
components such as circulators, inductors, and isolators, but these are bulky and can-
not be integrated with miniaturized electronic circuitry. This limits the utility of the
magnetic components as well as their ability to impact overall system performance
and function. Reducing the size, weight, and power (SWaP) of magnetic components
and integrating them onto semiconductor chips, however, could enable broader ex-
ploitation of magnetic materials and provide new mechanisms for the control and ma-
nipulation of EM signals. For instance, tighter integration of electronic and magnetic
components could yield smaller radar systems, higher bandwidth communication over
longer ranges, improved jam resistance, and more resilient EW systems. The M3IC
program is divided into three technical areas: integration of magnetic materials and
systems with semiconductor technology; accurate and efficient modeling of magnetic
phenomena from the molecular to the component system level; and exploitation of
magnetic phenomena in innovative component designs relevant to EM systems that
are important to DoD.

81
Active Interpretation of Disparate
Alternatives (AIDA)
Dr. William Corvey

https://www.darpa.mil/program/active-interpretation-of-disparate-alternatives.html

GPT-3 Summary: The goal of the AIDA program is to develop a multihypothesis

semantic engine that can generate alternative interpretations of events from multiple
unstructured sources, for use in noisy, conflicting, and potentially deceptive informa-
tion environments.

The technology methods for this will involve mapping knowledge elements
automatically derived from multiple media sources into a common semantic rep-
resentation, aggregating information derived from those sources, and generating
and exploring multiple hypotheses about the events, situations, and trends of interest.

If successful, this could provide a more comprehensive and accurate under-

standing of events happening around the world, and help to overcome the challenges
of independently

Full Details: The United States Government has an interest in developing and
maintaining a strategic understanding of events, situations, and trends around the
world, in a variety of domains. The information used in developing this understand-
ing comes from many disparate sources, in a variety of genres, and data types, and
as a mixture of structured and unstructured data. Unstructured data can include
text or speech in English and a variety of other languages, as well as images, videos,
and other sensor information. Even structured sources can vary in the expressiveness,
semantics, and specificity of their representations. Moreover, analysis is complicated
by the need to overcome the noisy, conflicting, and potentially intentionally decep-
tive nature of the data.It is a challenge for those who strive to achieve and maintain
an understanding of these events, situations, and trends that information from each
medium is often analyzed independently, without the context provided by information
from other media. Often, each independent analysis results in only one interpretation,
with alternatives being eliminated due to lack of evidence, even in the absence of con-
tradictory evidence. When these independent, impoverished analyses are combined,
generally late in the analysis process, the result can be a single apparent consen-
sus view that does not reflect a true consensus.The goal of Active Interpretation of
Disparate Alternatives (AIDA) is to develop a multihypothesis semantic engine that

82
generates explicit alternative interpretations of events, situations, and trends from a
variety of unstructured sources, for use in noisy, conflicting, and potentially decep-
tive information environments. This engine must be capable of mapping knowledge
elements automatically derived from multiple media sources into a common semantic
representation, aggregating information derived from those sources, and generating
and exploring multiple hypotheses about the events, situations, and trends of inter-
est. This engine must establish confidence measures for the derived knowledge and
hypotheses, based on the accuracy of the analysis and the coherence of the seman-
tic representation of each hypothesis. This engine must also be capable of utilizing
knowledge in the common semantic representation and the generated hypotheses as
alternate contexts for the media analysis algorithms by altering their models or prior
probabilities to enhance accuracy and resolve ambiguities in line with expectations
from the context. In addition, the engine must be able to communicate with its user
to reveal the generated hypotheses and to allow the user to alter the hypotheses or
to suggest new ones.

83
ReSource
Dr. Seth Cohen

https://www.darpa.mil/program/resource.html

GPT-3 Summary: The DARPA ReSource program’s goal is to enable the military
to quickly and easily procure critical supplies on the battlefield by engineering self-
contained, integrated systems that rapidly produce large quantities of supplies from
feedstock collected on-site. The program’s technology methods involve developing
software to pre-treat and breakdown mixed waste, including recalcitrant carbon-rich
polymers like those in common plastics, and reform and assemble upgradeable organic
molecules into strategic materials and chemicals. The program’s impact is that, if
successful, warfighters will be able to turn forage wastes into lubricants and nutrition,
supporting independent expeditionary

Full Details: Military logistic support has a large human cost in contested en-
vironments with no ability to create valuable materials when and where needed.
DARPA’s ReSource program aims to revolutionize how the military procures criti-
cal supplies on the battlefield by engineering self-contained, integrated systems that
rapidly produce large quantities of supplies from feedstock collected on-site. En-
visioned on-demand products include lubricants, adhesives, tactical fibers, potable
water, and edible macronutrients. Developed technologies will function simply, reli-
ably, and needing limited servicing in isolated environments. If successful, warfighters
will be able to turn forage wastes into lubricants and nutrition, supporting indepen-
dent expeditionary units and Humanitarian Assistance and Disaster Relief (HADR)
stabilization operations. ReSource performers will work toward producing a system
that will pre-treat and breakdown mixed waste, including recalcitrant carbon-rich
polymers like those in common plastics, reform and assemble upgradeable organic
molecules into strategic materials and chemicals, and recover purified usable prod-
ucts. The goal is to provide a versatile, operationally relevant, durable system that
converts single-use packaging, and other waste, into strategic and critical materials.
ReSource will pursue technologies that support this DoD objective by delivering a
cross-scale approach for warfighter support extending mission duration and expand-
ing operational flexibility.

84
Rational Integrated Design of Ener-
getics (RIDE)
Dr. Bartlett Russell

https://www.darpa.mil/program/rational-integrated-design-of-energetics.html

GPT-3 Summary: The RIDE program’s goal is to develop tools to speed up and
systematize energetics research, with the ultimate goal of enabling advanced develop-
ment and implementation of automation and discovery tools in the energetics domain.
Specifically, RIDE will develop 1) safe, semi-automated experimental capabilities that
integrate energetics ingredient synthesis with formulation development and testing;
and 2) theoretical, experimental, and/or statistical methods that enable safe, accurate
evaluation of key energetics properties at reduced scale.

Full Details: Automation and artificial intelligence are revolutionizing discovery and
production of functional molecules by enabling fast, reproducible experimentation and
efficient property optimization. These capabilities have already made a significant
impact on prevalent molecular classes, such as pharmaceuticals, but niche areas char-
acterized by unique chemical space, limited literature precedence, and requirements
for specialized experimental hardware have experienced relatively slow improvement.
One such area, critical to national security, is energetics.The DARPA Rational In-
tegrated Design of Energetics (RIDE) program will develop tools that speed and
systematize energetics research, enabling advanced development and implementation
of automation and discovery tools to the energetics domain. Specifically, these will
include: 1) safe, semi-automated experimental capabilities that integrate energetics
ingredient synthesis with formulation development and testing; and 2) theoretical,
experimental, and/or statistical methods that enable safe, accurate evaluation of key
energetics properties at reduced scale.

85
Real Time Machine Learning (RTML)
Dr. Dev Palmer

https://www.darpa.mil/program/real-time-machine-learning.html

GPT-3 Summary: GOAL: The RTML program seeks to create no-human-in-the-

loop hardware generators and compilers to enable the fully automated creation of ML
Application-Specific Integrated Circuits (ASICs) from high-level source code.
METHODS: The RTML program is focused on building an end-to-end general pur-
pose compiler that can transform a high level ML framework into Verilog.
IMPACT: This will allow future engineers to rapidly develop and deploy large-scale
real-time machine learning systems with customized hardware that can execute in-
tensive ML algorithmic tasks on chip, without the

Full Details: Driven by the rapidly evolving national security threat landscape,
future defense systems will need access to low size, weight, and power (SWaP) ar-
tificial intelligence (AI) solutions that can rapidly transition from idea to practice.
In recent years, the ability to learn from large datasets has advanced significantly
due to increases in hardware performance, advances in machine learning (ML) al-
gorithms, and the availability of high quality open datasets. However, current ML
systems are generally trained prior to deployment and are not capable of adapting
to new datasets in the field, limiting real-time “learning.” Critical next-generation
defense systems, such as autonomous vehicles and arrays of sensors, will be deployed
in distributed settings where resources for exporting newly encountered data might
be scarce or unavailable. The Real Time Machine Learning (RTML) program seeks to
solve this problem by creating no-human-in-the-loop hardware generators and com-
pilers to enable the fully automated creation of ML Application-Specific Integrated
Circuits (ASICs) from high-level source code. The RTML program is focused on
building an end-to-end general purpose compiler that can transform a high level ML
framework into Verilog. This, in turn, may allow future engineers to rapidly develop
and deploy large-scale real-time machine learning systems with customized hardware
that can execute intensive ML algorithmic tasks on chip, without the need for external
computational resources.

86
Battlefield Medicine
Dr. Amy Jenkins

https://www.darpa.mil/program/battlefield-medicine.html

GPT-3 Summary: The Battlefield Medicine program’s goal is to improve mili-

tary readiness by overcoming logistical obstacles to manufacturing and delivery of
urgently needed pharmaceutical products. The program’s technology methods focus
on developing miniaturized device platforms and techniques for synthesizing multiple
small-molecule active pharmaceutical ingredients (APIs) and therapeutic proteins.
The program’s impact is to enable effective small-batch pharmaceutical production
that obviates the need for individual drug stockpiling, cold storage, and complex
logistics.

Full Details: The Battlefield Medicine program supports military readiness in far-
forward deployed settings by overcoming logistical obstacles to manufacturing and
delivery of urgently needed pharmaceutical products used to treat emerging threats.
Battlefield logistics are a challenge regardless of the mission. Adversaries, terrain, and
the environment all serve to complicate the process of delivering supplies to warfight-
ers. The current Department of Defense (DoD) approach to medical supply logistics
is limited in its reach to far-forward emergency settings, response to emergent in-
theater threats, and utility for bio-preparedness stockpiling. It can often take weeks
to months to manufacture and airlift organic pharmaceuticals and protein therapeu-
tics to battlefield frontlines, meaning that critical medical supplies often do not arrive
in time where they are needed most. Furthermore, the need to prepare medical sup-
plies in advance based on an anticipated, specific threat can result in wasted materials,
labor, and money when that threat is not realized. The DoD needs a new approach.
Battlefield Medicine seeks to address this capability gap through two integrated re-
search thrusts: the Pharmacy on Demand (PoD) and Biologically-derived Medicines
on Demand (Bio-MOD) initiatives. The combined efforts seek to develop miniatur-
ized device platforms and techniques that can produce multiple small-molecule active
pharmaceutical ingredients (APIs) and therapeutic proteins in response to specific
battlefield threats and medical needs as they arise. PoD research is aimed at de-
veloping and demonstrating the capability to manufacture multiple APIs of varying
chemical complexity using shelf-stable precursors, while Bio-MOD research is focused
on developing novel, flexible methodologies for genetic engineering and modification
of microbial strains, mammalian cell lines, and cell-free systems to synthesize multiple
protein-based therapeutics. As a proof of concept, both PoD and Bio-MOD efforts
will seek to develop platforms for manufacturing single-dose levels of FDA-approved

87
APIs and biologics and demonstrate high purity, efficacy, and potency in short time-
frames. In developing a flexible, miniaturized synthesis and manufacturing platform,
Battlefield Medicine will leverage continuous flow approaches that will, if successful,
pave the path forward for enabling distributed, on-demand medicine manufacturing
capabilities in battlefield and other austere environments. Additionally, the platform
would have built-in flexibility to produce multiple types of therapeutics through its
modular reaction design. The ultimate vision for Battlefield Medicine is to enable
effective small-batch pharmaceutical production that obviates the need for individual
drug stockpiling, cold storage, and complex logistics.

88
Advanced Concept Compact Elec-
tron Linear-accelerator (ACCEL)
Dr. Matthew Higgins

https://www.darpa.mil/program/advanced-concept-compact-electron-linear-accelerator.html

GPT-3 Summary: GOAL: Develop a compact electron linear accelerator (LINAC)

for military applications.
TECHNOLOGY METHODS: Develop a LINAC that can create electron beams with
energy up to 35 MeV in a package that weighs less than 75 kg.
IMPACT: Provide multiple defense and homeland security benefits, such as enabling
portable medical treatments, remote detonation of improvised explosive devices, and
mobile imaging or inspection of shipping containers.

Full Details: Linear accelerators, LINACs for short, are devices that accelerate
electrons or other sub-atomic particles along a straight line to generate a beam of
high energy. LINACs have a variety of commercial uses such as generating X-rays
for cargo inspection, medical diagnostics, food sterilization, and even enabling precise
external radiation treatments to destroy cancer cells without damaging surrounding
tissue. To generate more powerful electron beams using current technology, however,
requires building larger LINACs that can grow to dozens of meters or longer depend-
ing on the application. Unfortunately, powerful LINACs are too large and heavy to
be practical for military use in the field.The Advanced Concept Compact Electron
Linear-accelerator (ACCEL) program’s goal is to develop a powerful, deployable elec-
tron LINAC for military applications.A high-power compact, rugged accelerator that
can be transported by truck or aircraft to austere locations would provide multiple
defense and homeland security benefits. It could be used for medical treatments in
locales without advanced hospitals, remote detonation of improvised explosive devices
(IEDs), and mobile imaging or inspection of shipping containers’ contents to counter
chem-bio and radiological threats. A deployable LINAC could also enable portable
sterilization for foods and surfaces to prevent contamination and infection in deployed
environments.The program seeks to develop a compact LINAC that can create elec-
tron beams with energy up to 35 mega-electron volts (MeV) in a package that weighs
less than 75 kilograms and can fit in a 1-meter long cylinder with 0.4-meter diameter.
The LINAC will need to be ruggedized so that it can be transported by air, rail, or
truck and operate in widely varying heat or cold conditions.

89
Personalized Protective Biosystem
(PPB)
Dr. Eric Van Gieson

https://www.darpa.mil/program/personalized-protective-biosystem.html

GPT-3 Summary: The Personalized Protective Biosystem (PPB) program seeks to

reduce the need for protective equipment while increasing protection against chemical
and biological (CB) threats. This will be achieved with lightweight materials and
adaptable, tissue-protective countermeasures acting independently, or as an ensemble,
to provide on-demand, broad spectrum, and rapid protection. The program will
leverage molecular components or commensal organisms at key points of vulnerability
to remove protective equipment burden from the user. Successful PPB technologies
would therefore change how the military and public health communities perform in
unpredictable threat environments.

Full Details: Chemical and biological (CB) threats have become increasingly ubiqui-
tous and diverse. They present a risk to our stability operators in pandemic outbreak
scenarios, and our warfighters serving in diverse operating environments. State-of-
the-art protective equipment continues to severely limit mobility and performance of
the user because of a dependence upon protective garments that are thick, heavy, and
cumbersome. The Personalized Protective Biosystem (PPB) program aims to reduce
protective equipment needs while increasing protection against existing and future
CB threats. PPB technologies aims to improve stability and provide flexibility for
field-forward individuals operating in austere environments–regardless of threat. This
will be achieved with lightweight materials and adaptable, tissue-protective counter-
measures acting independently, or as an ensemble, to provide on-demand, broad spec-
trum, and rapid protection. Most importantly, the program will leverage molecular
components or commensal organisms at key points of vulnerability to remove protec-
tive equipment burden from the user. Successful PPB technologies would therefore
change how the military and public health communities perform in unpredictable
threat environments.

90
Quantum Imaging of Vector Elec-
tromagnetic Radiation (QuIVER)
Dr. Jonathan Hoffman

https://www.darpa.mil/program/quantum-imaging-of-vector-electromagnetic-radiation.html

GPT-3 Summary: GOAL: Develop arrayed vector magnetometers to improve the

ability to locate and identify magnetic objects.
TECHNOLOGY METHODS: Use state-of-the-art magnetometers to measure the
Jacobian Tensor of the magnetic field.
IMPACT: Enable sensing from large distances and improve the ability to locate and
identify magnetic objects.

Full Details: The most common application of magnetic sensors is for determining
the location of magnetic objects. However, a single scalar magnetometer measurement
is not sufficient by itself in locating an object because many combinations of object
location and magnetic strength will produce the same sensor reading.Since the ob-
ject’s magnetic field decreases rapidly with distance (1/r3), and there is the constant
presence of the Earth’s large magnetic field that can overwhelm a signal, magnetome-
ters currently need to operate at close range. As a result, finding magnetic objects is
performed by methodically surveying an area, which requires a significant amount of
time.The Quantum Imaging of Vector Electromagnetic Radiation (QuIVER) program
aims to develop arrayed vector magnetometers from which one can measure spatial
derivatives of the vector field forming the Jacobian Tensor of the magnetic field. This
tensor will enable QuIVER to uniquely locate a magnetic object, its relative orienta-
tion, and its strength, with a single measurement. Knowledge of the tensor can also
enable the separation of the Earth’s magnetic field from that of the object of interest.
This ability coupled with the exquisite sensitivity of state-of-the-art magnetometers
will enable sensing from large distances.

91
Time-Aware Machine Intelligence (TAMI)
Lt. Col. C. David Lewis, USAF

https://www.darpa.mil/program/time-aware-machine-intelligence.html

GPT-3 Summary: GOAL: The TAMI program will develop new time-aware neural
network architectures that introduce a meta-learning capability into machine learning.

METHODS: Leverage the latest research on meta-learning in neural networks

to develop and prototype concrete computational models.

IMPACT: Enable an AI system to develop a detailed self-understanding of the

time dimensions of its learned knowledge and eventually be able to ”think in and
about time” when exercising its learned task knowledge in task performance.

Full Details: The Time-Aware Machine Intelligence (TAMI) program will develop
new time-aware neural network architectures that introduce a meta-learning capabil-
ity into machine learning.This meta-learning will enable a neural network to capture
the time-dependencies of its encoded knowledge. TAMI’s vision is for an AI sys-
tem to develop a detailed self-understanding of the time dimensions of its learned
knowledge and eventually be able to ”think in and about time” when exercising its
learned task knowledge in task performance.TAMI draws inspiration from ongoing
research in time-processing mechanisms in human brains. A large number of compu-
tational models have been introduced in computational neuroscience to explain time
perception mechanisms in the brain. TAMI will go a step further to develop and
prototype concrete computational models. TAMI will leverage the latest research on
meta-learning in neural networks.

92
Hybrid AI to Protect Integrity of
Open Source Code (SocialCyber)
Dr. Sergey Bratus

https://www.darpa.mil/program/hybrid-ai-to-protect-integrity-of-open-source-code.html

GPT-3 Summary: GOAL: The goal of this program is to protect the integrity of
open source code used by the Department of Defense.

METHODS: This program will use hybrid methods to combine analyses of

source code, development-related communication artifacts, and multi-modal social
media activities related to open source development.

IMPACT: This program will help to protect the integrity of critical open source
infrastructure used by the Department of Defense.

Full Details: The Department of Defense (DoD) has critical dependencies on open
source software (OSS) throughout its supply chain, including operating systems, vir-
tualization systems, and hypervisors as well as tool chains for software development.
The DoD’s use of OSS saves cost, increases maintainability, and attracts developer
talent, but also creates an unprecedented attack surface, in which many trusted soft-
ware parts and paths are exposed to hostile manipulation. Manipulators can leverage
the full scope of social mechanisms and incentives that make the OSS sociotechnical
ecosystem so valuable.Today, the integrity of these ecosystems relies on manual ef-
fort by their respective stewards, who typically act on implicit trust and perceived
reputation. Situational awareness is a matter of heuristics, such as the so-called “bus
factor” of a project, and hunches based on long-term community participation. As
a result, OSS stewards struggle to protect the integrity of their projects. Moreover,
tracking the dependencies of modern software ecosystems requires sustained, dedi-
cated effort that cannot be expected of any individual or group of stewards.For OSS
projects that take the stance of not publicly distinguishing between exploitable bugs
and functional bugs, adversaries may glean critical information before mitigations are
completed, and interfere with the mitigations. Social media campaigns to disrupt or
distract OSS developer communities can be highly effective, even if mounted by a
few individuals against a large and well-established community. Since development
of OSS is essentially a social process, such disruptions present a growing concern.The
AIE SocialCyber Opportunity aims to explore capabilities to detect and counteract
cyber-social operations that may target OSS developer communities. SocialCyber

93
seeks to explore hybrid methods that combine analyses of source code, development-
related communication artifacts, and multi-modal social media activities related to
open source development to protect the integrity of open source infrastructure critical
to the DoD.Additional information is available in the Hybrid AI to Protect Integrity
of Open Source Code (SocialCyber) Program Announcement.

94
Cooperative Secure Learning (CSL)
Dr. Joshua Baron

https://www.darpa.mil/program/cooperative-secure-learning.html

GPT-3 Summary: GOAL: The Cooperative Secure Learning (CSL) program will
develop methods to protect data, models, and model outputs among a community
of entities desiring to securely share their information to better inform ML model
development.
METHODS: The program will focus on developing working prototypes of computa-
tional techniques for improving ML models, and providing insights and methods that
support privacy preservation and data security. Underlying algorithms will be evalu-
ated based on their accuracy and privacy as well as their computational feasibility.
IMPACT: The CSL effort will significantly expand multi-organization ML capabili-
ties, leading to the construction of better informed and more robust

Full Details: Machine Learning (ML) requires vast amounts of data, but often the
data sets that enrich the models are owned by different parties and protected by
privacy, security, trade secrets, or regulatory requirements. Likewise, the applied ML
models (e.g., classifiers) are often owned by different parties and may be proprietary,
requiring stringent protection to reduce the threat of exposure for the input data
and modeling results. Due to these limitations, organizations in the government and
private sector are unable to cooperate fully in model training and development to
gain the best performance of ML systems.The Cooperative Secure Learning (CSL)
effort aims to develop methods to protect data, models, and model outputs among
a community of entities desiring to securely share their information to better inform
ML model development. CSL seeks to enable multiple parties to cooperate for the
purpose of improving each other’s ML models while assuring that each entity’s indi-
vidual, pre-existing datasets and models will remain private. This effort will focus on
developing working prototypes of computational techniques for improving ML mod-
els, and provide insights and methods that support privacy preservation and data
security. Underlying algorithms will be evaluated based on their accuracy and pri-
vacy as well as their computational feasibility. Possible technical approaches can
draw upon cryptographic methods (e.g., secure multiparty computation, homomor-
phic encryption, etc.,) differential privacy, and other methodologies. If successful, the
CSL effort will significantly expand multi-organization ML capabilities, leading to the
construction of better informed and more robust ML models without compromising
privacy.Additional information is available in the CSL Program Announcement.

95
World Modelers
Dr. Joshua Elliott

https://www.darpa.mil/program/world-modelers.html

GPT-3 Summary: GOAL: The World Modelers program will develop technology
to provide a comprehensive understanding of complicated, dynamic national security
questions.

TECHNOLOGY METHODS: The technology will integrate qualitative causal

analyses with quantitative models and relevant data.

IMPACT: The technology will provide parameterized, quantitative projections

within weeks or even hours, compared to the months or years it takes today to
understand simpler systems.

Full Details: The World Modelers program aims to develop technology that in-
tegrates qualitative causal analyses with quantitative models and relevant data to
provide a comprehensive understanding of complicated, dynamic national security
questions. The goal is to develop approaches that can accommodate and integrate
dozens of contributing models connected by thousands of pathways—orders of mag-
nitude beyond what is possible today. World Modelers analyses are intended to be
timely enough to recommend specific actions that could avert crises. The program
seeks to develop technologies that will provide clearly parameterized, quantitative
projections within weeks or even hours of processing, compared to the months or
years it takes today to understand considerably simpler systems. World Modelers
technologies will be applied to increasingly varied use cases as they mature through
the phases of the program. Questions for analysis will typically be framed at sub-
national scales and look one to five years into the future, although the factors that
influence outcomes of interest might operate on larger spatial and temporal scales.
This subnational focus reflects the changing nature of conflict and security, which,
increasingly, plays out in cities and districts. The first use case of World Modelers is
food insecurity resulting from the interactions of multiple factors, including climate,
water availability, soil viability, market instability, and physical security.

96
Autonomous Diagnostics to Enable
Prevention and Therapeutics (ADEPT)
Dr. Amy Jenkins

https://www.darpa.mil/program/autonomous-diagnostics-to-enable-prevention-and-therapeutics.htm

GPT-3 Summary: The ADEPT program’s goal is to develop technologies to rapidly

identify and respond to threats posed by natural and engineered diseases and tox-
ins. The program’s four main technology thrusts are simple-to-use, on-demand di-
agnostics; novel methods for rapidly manufacturing new vaccines; novel tools for
targeted drug delivery and in vivo diagnostics; and novel methods for imparting
near-immediate immunity.

Full Details: The Autonomous Diagnostics to Enable Prevention and Therapeutics

(ADEPT) program supports individual troop readiness and total force health protec-
tion by developing technologies to rapidly identify and respond to threats posed by
natural and engineered diseases and toxins. A subset of ADEPT technologies specif-
ically support use by personnel with minimal medical training, delivering centralized
laboratory capabilities even in the low-resource environments typical of many mili-
tary operations. The program is part of a portfolio of DARPA-funded research aimed
at providing options for preempting or mitigating constantly evolving infectious dis-
ease threats. The ADEPT program’s four thrusts cover simple-to-use, on-demand
diagnostics for medical decision-making and accurate threat-tracking; novel meth-
ods for rapidly manufacturing new types of vaccines with increased potency; novel
tools to engineer mammalian cells for targeted drug delivery and in vivo diagnostics;
and novel methods to impart near-immediate immunity to an individual using anti-
bodies. ADEPT has pioneered use of nucleic-acid-based anti-infective technologies,
valuable for their efficacy and adaptability. These tools—primarily coded genetic in-
structions to the body on how to produce its own protective antibodies against a spe-
cific threat—have the advantages of being easily manufactured at scale using largely
synthetic processes, transported and stored without many of the cold-chain logis-
tics required by traditional medical countermeasures, delivered with near-immediate
efficacy, and safely expressed in the body for only a limited duration, causing no
permanent alteration to the genome.

97
Environmental Microbes as a Bio-
Engineering Resource (EMBER)
Dr. Linda Chrisey

https://www.darpa.mil/program/environmental-microbes-as-a-bioengineering-resource.html

GPT-3 Summary: The goal of the EMBER program is to develop a biotechnology-

based separation and purification strategy for Rare Earth Elements (REEs) from
under-utilized domestic sources.

The technology methods of the program include the separation of REE mix-
tures into individual elements using aqueous processes, inter-conversion of REE
salts/oxides to facilitate production of manufacturing-ready forms, and new assays
for high-throughput analysis of REE-containing cells and biomolecules.

The impact of the program is to deliver multiple capabilities such as the sepa-
ration of REE mixtures

Full Details: The Environmental Microbes as a BioEngineering Resource (EMBER)

program will develop a biotechnology-based separation and purification strategy for
Rare Earth Elements (REEs) from under-utilized domestic sources such as phosphate
mine waste, acid mine drainage, and electronics recycling processes.The program aims
to deliver multiple capabilities such as the separation of REE mixtures into individual
elements using aqueous processes; inter-conversion of REE salts/oxides to facilitate
production of manufacturing-ready forms (e.g., halides, phosphates, nitrates); and
new assays for high-throughput analysis of REE-containing cells and biomolecules.

98
Tactical Boost Glide (TBG)
Dr. Peter ErblandLt. Col. Joshua Stults

https://www.darpa.mil/program/tactical-boost-glide.html

GPT-3 Summary: The Tactical Boost Glide (TBG) program is a DARPA/U.S. Air
Force effort to develop and demonstrate technologies for air-launched, tactical-range
hypersonic boost glide systems. The TBG program plans to focus on three primary
objectives: Vehicle Feasibility, Effectiveness, and Affordability. The program is using
a disciplined systems engineering approach to define demonstration system objectives
and identify enabling technologies needed for future systems. The TBG program is
exploiting the technical knowledge and lessons derived from development and flight
testing of previous boost glide systems, including the Hypersonic Technology Vehicle
2(

Full Details: Systems that operate at hypersonic speeds—five times the speed of
sound (Mach 5) and beyond—offer the potential for military operations from longer
ranges with shorter response times and enhanced effectiveness compared to current
military systems. Such systems could provide significant payoff for future U.S. offen-
sive strike operations, particularly as adversaries’ capabilities advance. The Tactical
Boost Glide (TBG) program is a joint DARPA/U.S. Air Force (USAF) effort that
aims to develop and demonstrate technologies to enable future air-launched, tactical-
range hypersonic boost glide systems. In a boost glide system, a rocket accelerates
its payload to high speeds. The payload then separates from the rocket and glides
unpowered to its destination. The TBG program plans to focus on three primary ob-
jectives: Vehicle Feasibility—Vehicle concepts possessing the required aerodynamic
and aerothermal performance, controllability and robustness for a wide operational
envelope Effectiveness—System attributes and subsystems required to be effective in
relevant operational environments Affordability—Approaches to reducing cost and
increasing value for both the demonstration system and future operational systems
TBG is a two-phase effort that plans to include ground and flight testing to mature
critical technologies, and aims to demonstrate the system performance achievable
through the integration of those technologies. The program is using a disciplined
systems engineering approach to define demonstration system objectives and identify
enabling technologies needed for future systems. The TBG program is exploiting the
technical knowledge and lessons derived from development and flight testing of pre-
vious boost glide systems, including the Hypersonic Technology Vehicle 2 (HTV-2).

99
Secure Handhelds on Assured Re-
silient networks at the tactical Edge
(SHARE)
Dr. Mary R. Schurgot

https://www.darpa.mil/program/secure-handhelds-on-assured-resilient-networks-at-the-
tactical-edge.html

GPT-3 Summary: The SHARE program is designed to develop handheld devices

that can securely exchange data at different security levels. The program will test the
devices in various scenarios and then transition the program to the military branches
in 2021.

Full Details: The security of communications and information for troops operating
in tactical warfighting environments is limited. At the tactical edge, end-to-end con-
nections on secure servers are typically unreliable, negatively impacting operations
and coordination with coalition partners. Additionally, current infrastructure require-
ments for such secure connections can be large and cumbersome, limiting tactical
deployment in fast-moving operations. To overcome these issues, Dr. Joseph Evans
(the original Program Manager) and now, Dr. Mary Schurgot, from the Strategic
Technology Office at DARPA, are leading the Secure Handhelds on Assured Resilient
networks at the tactical Edge (SHARE) program, which is developing handheld de-
vices that can reliably exchange data at varying security levels. “Troops forward
deployed today have to have multiple laptops or devices that are approved to com-
municate at various levels of classification,” said Dr. Evans, former DARPA program
manager. “The vision of SHARE is to develop software that moves the management
functions of multiple security levels from a handful of data centers down to trusted,
handheld devices on the tactical edge.” After successful testing in June 2019 at the
Asymmetric Warfare Group urban training facility at Fort A.P. Hill, VA, DARPA
has started phase three of the SHARE program. As part of further development of
SHARE security and networking capabilities, a series of test and integration events
are planned throughout the 2020 – 2021 timeframe. Test participants include the
75th Ranger Regiment and U.S. Army Medical Research and Development Com-
mand Telemedicine and Advanced Technology Research Center, both of whom have
unique concepts of employment that will stress the viability of SHARE to support
operational deployment. It is projected the program will transition to the military
branches in 2021.

100
Cyber-Hunting at Scale (CHASE)
Mr. Tejas Patel

https://www.darpa.mil/program/cyber-hunting-at-scale.html

GPT-3 Summary: The CHASE program seeks to develop automated tools to

detect and characterize novel attack vectors, collect the right contextual data, and
disseminate protective measures both within and across enterprises. CHASE will
focus on protective measures that a network owner has the authority to execute
within their own environment, as well as measuring the accuracy and efficiency of
threat detection techniques.

Full Details: Networks within the United States and abroad face increasingly broad-
spectrum cyber threats from numerous actors and novel attack vectors. Malicious ac-
tivity also crosscuts organizational boundaries, as nefarious actors use networks with
less protection to pivot into networks containing key assets. Detection of these threats
requires adjustments to network and host sensors at machine speed. Additionally, the
data required to detect these threats may be distributed across devices and networks.
In all of these cases, the threat actors are using technology to perpetrate their at-
tacks and hide their activities and movement, both physical and virtual, inside DoD,
commercial, and Internet Access Provider (IAP) networks. Enterprise-sized networks
present challenges in terms of both their size and distributed structure. Today’s
state-of-the-art commercial tools do not directly address the scale and speed needed
to provide the best defense for multiple networks. Networks lack robust mechanisms
to collect, share, and respond to threat intelligence. Data required to detect and char-
acterize malicious activities may be diffused and may be located across network and
endpoint devices. Further, cyber-relevant data (including data that may contain in-
formation useful for detection and characterization) routinely exceeds total available
storage, bandwidth, and analysis capability, often by several orders of magnitude.
Of data that is able to be stored, only some is currently analyzed, and of all alerts
generated, only a fraction are threat related. Storage and processing limitations
abound, so cyber defenders require tools that strategically direct resources toward
the data that actually contains information about threats. Current commercially
available tools may output thousands of alerts and false positives per day that often
cannot be verified due to a lack of processing capacity. Static data retention policies
sometimes result in the deletion of relevant data prior to investigation. Additionally,
current tools may neither proactively detect novel attack vectors nor detect coordi-
nated attacks distributed across multiple organizations. Traditionally, cyber defense
technologies focus predominantly upon either host data or network data. Malicious

101
activity, however, crosscuts networks and hosts. Real-time detection of threats within
or across very large enterprise networks is not simply an issue of scale, but also a chal-
lenge due to the variable nature of malicious activities and their presentations. The
CHASE program seeks to develop automated tools to detect and characterize novel
attack vectors, collect the right contextual data, and disseminate protective measures
both within and across enterprises. CHASE aims to prototype components that en-
able network owners to reconfigure sensors and disseminate protective measures at
machine speed with appropriate levels of human supervision. CHASE technologies
will explore real-time investigations of potential cyber threats through adaptive data
collection. Threat detection algorithms developed under CHASE may be tailored to
characterize and react to specific classes of threats in the context of different data
types and data sources. Additionally, these algorithms may work in concert to de-
termine probabilities of the reality of threats, as well as indicate requirements for
additional data that should be collected. As such, the goal of CHASE is to develop
foundational technologies for detection, characterization, and strategic data manage-
ment. Enhanced threat detection may cue the generation of automated protective
measures. CHASE will focus on protective measures that a network owner has the
authority to execute within their own environment, as well as measuring the accuracy
and efficiency of threat detection techniques.

102
Traveling-Wave Energy Enhancement
Devices (TWEED)
Dr. David K. Abe

https://www.darpa.mil/program/traveling-wave-energy-enhancement-devices.html

GPT-3 Summary: The goal of the TWEED program is to develop and demonstrate
new approaches to active gain in sub-THz/THz, acoustoelectric, or spin wave devices.
The methods used will be based on coupled- or traveling-wave interaction. The impact
of this program will be the development of new devices with potential applications in
THz electronics, signal processing, and magnonics.

Full Details: Amplification based on the synchronous interaction between an elec-

tromagnetic traveling wave and a monoenergetic stream of electrons is a well-known
phenomenon in the field of vacuum electronics and commonly used in devices such
as traveling-wave tubes (TWTs). Active gain through traveling wave interaction has
been observed or postulated in other domains but has yet to be widely exploited.The
benefits of coupled wave interaction are not limited to THz electronics. Acoustoelec-
tric amplification was first demonstrated in the 1960s. Since then, new piezoelectric
materials, device architectures, and fabrication techniques present opportunities to
create surface acoustic wave devices with active gain, with advantages ranging from
reduced insertion loss to net-positive signal amplification. Spin waves have found
application in devices such as frequency selective limiters, phase shifters, delay lines,
etc. and more recently are under consideration for nanoscale transmission and pro-
cessing of information. New techniques using phenomena such as spin-transfer torque
and spin-orbit torque show promise to produce spin wave amplification to overcome
conversion and propagation losses.The Traveling-Wave Energy Enhancement Devices
(TWEED) Microsystems Exploration (µE) topic aims to develop and demonstrate
new approaches using coupled- or traveling-wave interaction to achieve active gain
in sub-THz/THz, acoustoelectric, or spin wave devices. The goal of TWEED is to
demonstrate proof-of-concept devices that establish the viability of novel approaches
that can lead to revolutionary advances in areas such as THz electronics, signal pro-
cessing, and magnonics. Further, to support potential transitions, the µE emphasizes
device architectures and materials that are compatible with current commercial fab-
rication processes.

103
Modeling Influence Pathways (MIPs)
Dr. Brian Kettler

https://www.darpa.mil/program/modeling-influence-pathways.html

GPT-3 Summary: The goal of MIPs is to map and model the pathways that
information takes as it flows across online and offline platforms. This will be done by
connecting various identified influence messaging flows across platforms and learning,
mapping, and modeling which pathways are used by what types of information. The
hope is that this will help discover patterns that characterize these pathways and
enable earlier anticipation and detection of new influence messaging flows.

Full Details: Influence pathways are the dynamic means by which coordinated
influence messaging flows across online and traditional broadcast (“offline”) media
platforms and communities. For example, in the 1980s, disinformation that AIDS
was a U.S. bioweapon spread from less reputable newspapers to more reputable
ones as interest in the story grew. Similar narratives emerged and propagated at
a much larger scale and speed during the Ebola and COVID pandemics of 2014
and 2019, respectively, over newspapers, broadcast, and social media. False nar-
ratives reemerged in early 2022, with a variety of actors (e.g. foreign states and
fringe groups) propagating the false information across a number of online and of-
fline platforms and communities. Understanding the pathways by which different
types of information propagate across the information ecosystem is important. In
addition, identifying patterns that emerge among these pathways, such as the prop-
agation from certain niche platforms to more mainstream platforms, is essential to
understanding influence operations. Modeling Influence Pathways (MIPs) explores
artificial intelligence (AI) technologies for the following:Connecting various identified
influence messaging flows across platformsLearning, mapping, and modeling which
pathways are used by what types of informationDiscovering patterns that charac-
terize these pathwaysMIPs complements maturing capabilities for identification of
misinformation, disinformation, and manipulated information with better mapping
and understanding of the pathways used to disseminate and amplify that informa-
tion. These pathways are not static but rather adapt to changes in the information
ecosystem in response to emerging platforms, increasing platform moderation or cen-
sorship, or new types of messaging. Today, pathway discovery is often a manually
intensive, retrospective process that fails to match the speed and scale of the infor-
mation ecosystem. More automated, accurate, and timely identification of influence
pathways will enable earlier anticipation and detection of new influence messaging
flows; and better forecasting of the impact of changes to the ecosystem (e.g., plat-

104
form policies) on information flows. The MIPS Program Announcement is available
at https://sam.gov/opp/a28c282ea87f42568492247671580d0a/view

105
Active Social Engineering Defense
(ASED)
Mr. Bernard McShea

https://www.darpa.mil/program/active-social-engineering-defense.html

GPT-3 Summary: GOAL: The ASED program will develop technology to auto-
matically identify, disrupt, and investigate social engineering attacks.
METHODS: The ASED technology will detect attacks, intervene in communications
between users and potential attackers, and coordinate investigations into the source
of the attacks.
IMPACT: The ASED technology will help protect against cyber attacks by making
it more difficult for attackers to exploit humans.

Full Details: Over the past 40 years, our world has become increasingly connected.
These connections have enabled major advances in national security from pervasive
real-time intelligence and communications to optimal logistics. With this connectivity
has come the threat of cyber attacks on both military systems and critical infrastruc-
ture. While we focus the vast majority of our security efforts on protecting computers
and networks, more than 80 percent of cyber attacks and over 70 percent of those
from nation states are initiated by exploiting humans rather than computer or net-
work security flaws. To build secure cyber systems, it is necessary to protect not only
the computers and networks that make up these systems but their human users as
well.We call attacks on humans “social engineering” because they manipulate or “en-
gineer” users into performing desired actions or divulging sensitive information. The
most general social engineering attacks simply attempt to get unsuspecting internet
users to click on malicious links. More focused attacks attempt to elicit sensitive in-
formation, such as passwords or private information from organizations or steal things
of value from particular individuals by earning unwarranted trust.These attacks al-
ways have an “ask,” a desired behavior that the attacker wants to induce from the
victim. To do this, they need trust from the victim, which is typically earned through
interaction or co-opted via a spoofed or stolen identity. Depending on the level of
sophistication, these attacks will go after individuals, organizations, or wide swathes
of the population.Social engineering attacks work because it is difficult for users to
verify each and every communication they receive. Moreover, verification requires
a level of technical expertise that most users lack. To compound the problem, the
number of users that have access to privileged information is often large, creating a
commensurately large attack surface.The Active Social Engineering Defense (ASED)

106
program aims to develop the core technology to enable the capability to automatically
identify, disrupt, and investigate social engineering attacks. If successful, the ASED
technology will do this by actively detecting attacks, intervening in communications
between users and potential attackers, and coordinating investigations into the source
of the attacks.

107
Ensuring Consistency of Systemic In-
formation (ECoSystemic)
Dr. Mark Flood

https://www.darpa.mil/program/ensuring-consistency-of-systemic-information.html

GPT-3 Summary: The ECoSystemic program will develop techniques for robust
recovery of federated financial information systems. This will be done by improving
the mutual consistency of federated backups along with processes for recovering a dis-
tributed system of potentially global scale. If successful, this will enable information
systems facing threats or outages to promptly return to a functional and mutually
consistent restored state.

Full Details: The Ensuring Consistency of Systemic Information (ECoSystemic)

program aims to develop innovative techniques for the robust recovery of federated
financial information systems. By improving the mutual consistency of federated
backups along with processes for recovering a distributed system of potentially global
scale, information systems facing threats or outages can promptly return to a func-
tional and mutually consistent restored state. This resilience and consistency in
federated data stores has broad applicability in both military and commercial do-
mains.ECoSystemic spotlights the U.S. financial sector as a use case because of its
strategic importance and extensive data federation. ECoSystemic will focus on (1)
modeling resilience within this global-scale distributed system, and (2) identifying
potential safeguards and responses to threats. If individual firms are vulnerable to a
common shock, there is nothing to prevent multiple key institutions from suffering an
outage simultaneously. ECoSystemic seeks innovative techniques to promote success-
ful, simultaneous recovery across diverse local backups, enabling mutual consistency
for a restored federated state.Additional information is available in the ECoSystemic
Program Announcement.

108
Oversight
Dr. Lael Rudd

https://www.darpa.mil/program/oversight.html

GPT-3 Summary: GOAL: Develop software to enable autonomous constant

custody of up to 1,000 targets from space assets.

METHODS: Leverage existing and/or state-of-the-art networks to provide col-

laboration between satellite and ground resources.

IMPACT: Improve tracking of targets in contested environments where resources

and targets may be highly dynamic.

Full Details: DARPA, the U.S. Space Force, and the Space Development Agency
(SDA) are developing new satellite constellations to increase the tactical capabili-
ties of U.S. space systems, however new methods are needed to effectively leverage
them. Oversight seeks software solutions to enable autonomous constant custody,
or knowledge of target location within accuracies necessary for mission needs, of up
to 1,000 targets from space assets through management of available satellite hard-
ware resources. The project aims to support both peacetime and wartime monitoring
of high value targets in contested environments where resources and targets may be
highly dynamic.Current practices require human operators for exquisite satellite solu-
tions. This arrangement does not scale well for the numbers of targets that Oversight
is considering. Reliance on individual ground station operators significantly increases
latency and minimizes tactical utility of satellite sensor data. Oversight will develop
the autonomy necessary to track targets with the operator overseeing at an aggre-
gate level. It will also leverage existing and/or state-of-the-art networks to provide
collaboration between satellite and ground resources.

109
Friend or Foe
Dr. Paul Sheehan

https://www.darpa.mil/program/friend-or-foe.html

GPT-3 Summary: The Friend or Foe program seeks to develop biosurveillance

technology that can detect bacterial pathogens as, or even before, they threaten the
military and homeland. The goal of the program is to quickly determine whether an
unknown bacterium is harmless or virulent by directly identifying pathogenic behav-
ior, avoiding conventional strategies that rely on known biomarkers.

Full Details: The Friend or Foe program aims to develop biosurveillance technology
that can detect bacterial pathogens as, or even before, they threaten the military and
homeland. The goal of the program is to quickly determine whether an unknown bac-
terium is harmless or virulent by directly identifying pathogenic behavior, avoiding
conventional strategies that rely on known biomarkers.The risk posed by unknown
bacteria is increasing as the global environment changes, populations expand, and
tools for genetic engineering proliferate. After new pathogens emerge, they can easily
spread due to increased global travel among dense urban centers. Any one of these
unfamiliar strains might pose a health risk to deployed service members—especially
among first responders to outbreaks of infectious disease. Current biosurveillance
strategies based on biochemical markers fall short in identifying potentially harm-
ful bacteria since they do not work on undiscovered bacterial strains or on bacteria
engineered to evade detection. To overcome this problem, Friend or Foe aims to
characterize bacteria by identifying pathogenic behavior itself. The envisioned, high-
throughput Friend or Foe platform seeks to screen many unfamiliar strains of bacte-
ria at once to reveal their phenotypes, first by extracting and isolating bacteria from
complex environmental samples, then by sustaining them in simulated host organism
environments, and finally by interrogating the bacteria to determine pathogenicity.
In effect, the systems will play the biological equivalent of the game “Twenty Ques-
tions,” subjecting bacteria to a battery of physical and chemical tests to determine
pathogenicity.This screening would flag dangerous bacteria for subsequent genetic
sequencing to map the newly discovered pathogenic traits to specific genes. Early
identification of pathogenic genes could accelerate research into the development of
therapeutic countermeasures or provide new assays for more conventional, front-line
biosurveillance and diagnostic platforms.

110
Space Environment Exploitation (SEE)
Lt. Col. C. David Lewis, USAF

https://www.darpa.mil/program/space-environment-exploitation.html

GPT-3 Summary: GOAL: The Space Environment Exploitation (SEE) program

will develop new models and sensing modalities to predict and observe the dynamics
of the near-earth space environment.

TECHNOLOGY METHODS: The SEE program explores how to go beyond

magnetohydrodynamic descriptions of the magnetosphere, ionosphere, thermosphere
coupled system to include wave/wave, wave/particle, and particle/particle interac-
tions while using the latest advances in high performance computing such as GPUs
and TPUs.

IMPACT: The expected outcomes of SEE will give future commanders and
operators the necessary

Full Details: The Space Environment Exploitation (SEE) program seeks to de-
velop new models and sensing modalities to predict and observe the dynamics of
the near-earth space environment. The SEE program explores how to go beyond
magnetohydrodynamic descriptions of the magnetosphere, ionosphere, thermosphere
coupled system to include wave/wave, wave/particle, and particle/particle interac-
tions while using the latest advances in high performance computing such as GPUs
and TPUs. Furthermore, SEE is exploring how to unify current space environmental
sensing networks to produce a common operating space environment picture and how
to develop low cost, non-traditional, exploitive, and expeditionary means to observe
near-earth plasma dynamics. Another big component of SEE is understanding the
viability of how Artificial Intelligence and Machine Learning can be used to help as-
similate environmental data into models and virtually produce synthetic data. The
expected outcomes of SEE will give future commanders and operators the necessary
and precise space environment situational awareness to make relevant space opera-
tional/tactical decisions and differentiate between human-made and natural dynamic
perturbations of the environment.

111
Heterogeneous Heterostructures (H2)
Dr. Thomas Kazior

https://www.darpa.mil/program/heterogeneous-heterostructures.html

GPT-3 Summary: GOAL: The H2 program seeks to develop new approaches

to creating low-defect interfaces between dissimilar materials in order to create
high-power density RF devices.

METHODS: Developing new methods for creating low-defect interfaces be-

tween dissimilar materials.

IMPACT: Creating RF transistors with a 10X increase in power density com-

pared to current state-of-the-art devices.

Full Details: Today’s state of the art (SoA) microwave and millimeter wave power
amplifiers have been made possible by the creation of lateral and vertical hetero-
junction semiconductor devices, such as high electron mobility transistors (HEMTs)
and heterojunction bipolar transistors (HBTs), respectively. To meet future mis-
sion requirements for longer range operation, wider bandwidth, and link robustness,
power amplifiers with higher output power than can be achieved with today’s SoA
are needed. This in turn will demand radio frequency (RF) device structures that
simultaneously exhibit higher charge density, higher breakdown field, good carrier
transport properties, and high thermal conductivity. Researchers have theorized that
potentially game changing device structures can be formed by integrating dissimilar
materials to solve current device performance limitations, but have not succeeded in
realizing these devices, primarily due to poor interface quality between dissimilar ma-
terials, which degrades the electronic properties.The Heterogeneous Heterostructures
(H2) Microsystems Exploration (µE) topic aims to create a new path for realizing
high power density, microwave, and millimeter wave devices by demonstrating het-
erogeneous heterojunctions. In particular, the H2 µE seeks to explore and develop
approaches to create low defect density interfaces between dissimilar materials. These
approaches will be used to create heterojunctions with low defect density and het-
erostructures that simultaneously exhibit high charge density, high breakdown field,
good carrier transport properties, and high thermal conductivity.The target outputs
will support the creation of RF transistors with a 10X increase in power density
compared with today’s SoA.

112
Enhanced Night Vision in Eyeglass
Form (ENVision)
Dr. Rohith Chandrasekar

https://www.darpa.mil/program/envision.html

GPT-3 Summary: The DARPA ENVision program will develop novel direct-view
NV systems that demonstrate an increased FOV and enhanced visual access across
infrared bands, all contained in a low-torque eyeglass form factor. ENVision will
also investigate the fundamental limits of low-input-intensity photon upconversion
bandwidth and efficiency. These new insights could potentially lead to further trans-
formative NV systems that require no optics or transduction, but simply directly
upconvert ambient infrared light into the visible.

Full Details: Despite decades of development, current night vision (NV) systems
are bulky and heavy, resulting in a large torque on the wearer’s neck. This torque
greatly limits the wearer’s agility and often leads to chronic injury over prolonged use.
Additionally, these systems provide the wearer with a narrow field of view (FOV)
and generally have limited spectral access to the near-infrared (NIR), greatly limiting
situational awareness. The DARPA ENVision program will leverage recent advances
in planar optics and transduction materials to develop novel direct-view NV systems
that demonstrate an increased FOV and enhanced visual access across infrared bands,
all contained in a low-torque eyeglass form factor. ENVision will also investigate
the fundamental limits of low-input-intensity photon upconversion bandwidth and
efficiency. These new insights could potentially lead to further transformative NV
systems that require no optics or transduction, but simply directly upconvert ambient
infrared light into the visible.

113
Accelerated Molecular Discovery (AMD)
Dr. Bartlett Russell

https://www.darpa.mil/program/accelerated-molecular-discovery.html

GPT-3 Summary: The goal of the Accelerated Molecular Discovery (AMD) pro-
gram is to develop new, systematic approaches to increase the pace of discovery and
optimization of high-performance molecules. The program is developing closed-loop
systems that exploit, build and integrate tools for: 1) extracting existing data from
databases and text; 2) executing autonomous experimental measurement and opti-
mization; and 3) incorporating computational approaches to develop physics-based
representations and predictive tools.

The AMD systems will provide a comprehensive computational and experi-

mental means to design, discover, validate and optimize new molecules, iteratively
and actively learning to more efficiently and effectively discover molecules that
enhance

Full Details: Efficient discovery and production of new molecules is essential to

realize capabilities across the DoD, from simulants and medicines essential to counter
emerging threats, to coatings, dyes and specialty fuels needed for advanced perfor-
mance. Current approaches to develop molecules for a given application are intuition-
driven, mired in slow iterative design/test cycles and ultimately limited by the specific
molecular expertise of the chemist, bottlenecks in the synthesis of candidates, and the
pace of by-hand validation and optimization of these candidates.DARPA’s Acceler-
ated Molecular Discovery (AMD) program is developing new, systematic approaches
that increase the pace of discovery and optimization of high-performance molecules.
Specifically, performers are developing closed-loop systems that exploit, build and
integrate tools for: 1) extracting existing data from databases and text; 2) execut-
ing autonomous experimental measurement and optimization; and 3) incorporating
computational approaches to develop physics-based representations and predictive
tools. AMD systems will provide a comprehensive computational and experimen-
tal means to design, discover, validate and optimize new molecules, iteratively and
actively learning to more efficiently and effectively discover molecules that enhance
performance in applications relevant to national security.

114
INTERfering and Co-Evolving Pre-
vention and Therapy (INTERCEPT)
Dr. Seth Cohen

https://www.darpa.mil/program/intercept.html

GPT-3 Summary: The goal of the INTERCEPT program is to create a new form
of medical countermeasure called therapeutic interfering particles (TIPs). These
particles are harmless, virus-derived particles with defective genomes that can only
replicate in the presence of virus. They compete with viral infection by competing
for essential viral components. TIPs are also susceptible to mutation over time and
co-evolve with the mutating virus, thus diminishing the virus’ ability to evade the
therapeutic.

The methods used in this program include novel molecular and genetic design
tools, high throughput genomic technologies, and advanced computational methods.

If

Full Details: Viral pathogens pose a continuous and shifting biological threat to
military readiness and national security overall in the form of infectious disease with
pandemic potential. Today’s limited vaccines and other antivirals are often circum-
vented by quickly mutating viruses that evolve to develop resistance to treatments
that are carefully formulated to act only specific strains of a virus. DARPA’s IN-
TERfering and Co-Evolving Prevention and Therapy (INTERCEPT) program aims
to harness viral evolution to create a novel, adaptive form of medical countermea-
sure—therapeutic interfering particles (TIPs)—that outcompetes viruses in the body
to prevent or treat infection. Whereas current preventive and therapeutic approaches
are designed to target viruses in their original state at the time of discovery or diag-
nosis, INTERCEPT uses viral evolution as the basis for its protective effect. Because
TIPs are harmless, virus-derived particles with defective genomes that can only repli-
cate in the presence of virus, they interfere with viral infection by competing for
essential viral components. And, just like their parent virus, TIPs are susceptible to
mutation over time and co-evolve with the mutating virus, thus diminishing the virus’
ability to evade the therapeutic.Over the course of this countermeasure-development
effort, INTERCEPT performer teams will use novel molecular and genetic design
tools, high throughput genomic technologies, and advanced computational methods
to address TIP safety, efficacy, long-term co-evolution, and generalizability. If success-

115
ful, INTERCEPT will deliver new treatments for fast-evolving viruses such as Ebola,
SARS, Dengue, Zika, and Chikungunya—providing broad coverage against multiple
strains—and make available a platform technology that could be readily adapted to
confront even engineered viral threats.

116
Seeker Cost Transformation (SECTR)
Dr. Paul Zablocky

https://www.darpa.mil/program/seeker-cost-transformation.html

GPT-3 Summary: GOAL: The SECTR program will develop new technologies
for air-launched weapons to improve their accuracy and function in GPS-denied
environments.

TECHNOLOGY METHODS: The program will develop new sensors and guid-
ance systems for weapons, as well as a new open architecture for the seeker.

IMPACT: The technologies developed under the program will transition to the
military Services to improve their capabilities.

Full Details: The Seeker Cost Transformation (SECTR) program seeks to de-
velop novel weapon terminal sensing and guidance technologies and systems for air-
launched, air-delivered weapons. SECTR technologies would enable weapons to ac-
quire fixed and moving targets with only minimal external support; achieve high
navigation accuracy in a GPS-denied environment; and be low size, weight, and cost.
The program aims to develop technologies and systems applicable to a wide range
of weapons and missions, such as small unit operations, suppression of enemy air
defenses, precision strike, and time-sensitive targets. The technical approach for the
sensing and processing hardware will embrace both passive electro-optical infrared
(EO/IR) sensors, which have evolved into very small and inexpensive devices in the
commercial market, and a reconfigurable processing architecture. The program also
seeks to develop a government-owned open architecture for the seeker with stan-
dardized interfaces between components (both hardware and software). Technologies
developed under this program would transition to the military Services.

117
Mobile Force Protection (MFP)
Dr. Gregory Avicola

https://www.darpa.mil/program/mobile-force-protection.html

GPT-3 Summary: GOAL: The MFP program will develop systems to thwart the
proliferation of small, unmanned aircraft systems.
TECHNOLOGY METHODS: The program includes three phases punctuated by
open-air demonstrations involving increasingly sophisticated threats and scenarios.
IMPACT: The program will culminate in a full-capability demonstration on a moving
vehicle or vessel by the end of Phase 3.

Full Details: DARPA’s Mobile Force Protection (MFP) program focuses on a chal-
lenge of increasing concern to the U.S. military: thwarting the proliferation of small,
unmanned aircraft systems. These systems – which include fixed- or rotary-wing air-
craft and have numerous advantages such as portability, low cost, commercial avail-
ability, and easy upgradeability – pose a fast-evolving array of dangers for U.S. ground
and maritime convoys. MFP aims to address threats from small unmanned aircraft
systems by developing scalable, modular, and affordable approaches that could be
deployed in the next few years and nimbly evolve with advances in threats, tactics,
and technology. The MFP program includes three phases punctuated by open-air
demonstrations involving increasingly sophisticated threats and scenarios. The pro-
gram intent is to culminate in a full-capability demonstration on a moving vehicle or
vessel by the end of Phase 3.

118
Atmospheric Water Extraction (AWE)
Dr. Seth Cohen

https://www.darpa.mil/program/atmospheric-water-extraction.html

GPT-3 Summary: The goal of the Atmospheric Water Extraction (AWE)

program is to provide potable freshwater for a range of military, stabilization, and
humanitarian needs.

The program will develop next-generation, scalable sorbent materials that can
rapidly extract water from ambient air. Additionally, the program will leverage
advanced modeling, innovative engineering, and additive manufacturing methods to
achieve a substantial reduction in the size, weight, and power (SWaP) requirements
compared to current atmospheric water generation technologies.

The AWE program will address water needs in two tracks: expeditionary and
stabilization. The expeditionary unit will provide sufficient

Full Details: Water transport is as mission-critical and as logistically challenging as

fuel transport for the U.S. military. Meeting deployed military water needs requires
equipment resources, consumes fuel, and endangers personnel. The goal of DARPA’s
Atmospheric Water Extraction (AWE) program is to provide potable freshwater for
a range of military, stabilization, and humanitarian needs through the development
of small, lightweight, low-powered, distributable systems that extract potable water
from the atmosphere to meet the drinking needs of individuals and groups, even in
extremely arid climates.AWE seeks to develop next-generation, scalable sorbent mate-
rials that can rapidly extract water from ambient air and leverage advanced modeling,
innovative engineering, and additive manufacturing methods to achieve a substantial
reduction in the size, weight, and power (SWaP) requirements compared to current
atmospheric water generation technologies. AWE will address water needs in two
tracks: expeditionary and stabilization. The expeditionary unit will provide suffi-
cient drinking water for an individual warfighter, with SWaP parameters restricted
by the need for portability and operation in austere environments. The stabilization
device will provide the daily drinking needs for up to ˜150 people (i.e., a company
or humanitarian mission), with SWaP requirements tailored to resources available to
missions of that scale.

119
Brandeis
Dr. Joshua Baron

https://www.darpa.mil/program/brandeis.html

GPT-3 Summary: The Brandeis program seeks to develop tools and techniques
that enable us to build systems in which private data may be used only for its
intended purpose and no other. The objective of Brandeis is to provide data owners
with mechanisms for protecting their data before sharing it with a data user. It will
also tackle a cognitive challenge: the volume and complexity of data means that
individuals or enterprises need a meaningful way to make choices about how to share
data, including understanding the implications of the use of any stored data about
them.

The potential impact of the Brandeis program is dramatic. Assured data pri-
vacy can

Full Details: The collection and analysis of information on massive scales has clear
benefits for society: it can help businesses optimize online commerce, medical workers
address public health issues and governments interrupt terrorist activities. Yet at the
same time, respect for privacy is a cornerstone principle of our democracy. The right
to privacy, as Louis Brandeis first expounded in 1890, is a consequence of modernity
because we better understand that harm comes in more ways than just the physical.
Numerous recent incidents involving the disclosure of data have heightened society’s
awareness of the extreme vulnerability of private information within cyberspace and
of the relationship of private data with personal and national security. There is a
growing desire to understand, control and manage the “digital contrail” of personal
information continually being produced – data that other people or organizations
could exploit. The Brandeis program seeks to develop the technical means to protect
the private and proprietary information of individuals and enterprises. The vision of
the Brandeis program is to break the tension between: (a) maintaining privacy and (b)
being able to tap into the huge value of data. Rather than having to balance between
them, Brandeis aims to build a third option – enabling safe and predictable sharing of
data in which privacy is preserved. The objective of Brandeis is to develop tools and
techniques that enable us to build systems in which private data may be used only
for its intended purpose and no other. It seeks to restructure our relationship with
data by providing the data owner with mechanisms for protecting their data before
sharing it with a data user. It will also tackle a cognitive challenge: the volume and
complexity of data means that individuals or enterprises need a meaningful way to

120
make choices about how to share data, including understanding the implications of
the use of any stored data about them.The potential impact of the Brandeis program
is dramatic. Assured data privacy can open the doors to personal medicine, effective
smart cities, detailed global data and fine-grained internet awareness.

121
Reconfigurable Imaging (ReImagine)
Dr. Whitney Mason

https://www.darpa.mil/program/reconfigurable-imaging.html

GPT-3 Summary: GOAL: The Reconfigurable Imaging (ReImagine) program will

develop a new approach to application development that is more similar to field
programmable gate array (FPGA)-based design than ASIC design.

TECHNOLOGY METHODS: 3-D integration will be used to interface with

virtually any type of imaging sensor. The mode of operation can be adapted either
through manual user control, through preset routines that can change many times
per second, or in response to context derived from the scene being observed.

IMPACT: The ReImagine ROICs could enable delivery of more actionable in-
formation

Full Details: Today’s imaging systems primarily perform a single or limited set
of measurements due, in large part, to the underlying readout integrated circuits
(ROICs), which sample the signal of interest and transfer the values off of the chip.
ROICs are currently designed for a specific mode of operation and are, effectively, ap-
plication specific integrated circuits (ASICs).The Reconfigurable Imaging (ReImag-
ine) program seeks to demonstrate that a software-reconfigurable imaging system is
capable of enabling revolutionary capabilities. The goals of the program are to create
a new approach to application development that is more similar to field programmable
gate array (FPGA)-based design than ASIC design, and develop the underlying the-
ory and algorithms that learn to collect the most valuable information when the
sensor can be configured for a variety of measurements. The ReImagine program
aims to demonstrate that a single, reconfigurable ROIC architecture can accommo-
date multiple modes of imaging operations that may be defined after a chip has been
designed.The ReImagine program takes advantage of 3-D integration, which makes
it possible to interface with virtually any type of imaging sensor (e.g. photodiode,
photoconductor, avalanche photodiode, or bolometer) and optimize it for any spec-
tral band (e.g. ultraviolet (UV) through very long-wave infrared (VLWIR)). More
importantly, this approach makes it possible to adapt the mode of operation either
through manual user control, through preset routines that can change many times per
second, or in response to context derived from the scene being observed. For example,
a single imager could present simultaneous regions of interest (ROIs) that can run at
different spatial or temporal resolutions or can run either active or passive.ReImagine

122
ROICs will also demonstrate that efficient computation within an ROI can enable
real-time analysis on much more complex scenes than traditional systems. The pro-
gram will build on this architecture to develop a concept of operation, application
requirements, modes of operation, and needed algorithms. The ReImagine ROICs
could enable delivery of more actionable information to the warfighter than has ever
been possible from a single imaging sensor.

123
Lifelong Learning Machines (L2M)
Mr. Ted Senator

https://www.darpa.mil/program/lifelong-learning-machines.html

GPT-3 Summary: GOAL: To develop AI systems that can learn and adapt like
humans.
TECHNOLOGY METHODS: Develop new architectures, mechanisms, and algo-
rithms based on discoveries in neuroscience and psychology.
IMPACT: Create AI systems that are safer and more efficient in a variety of tasks.

Full Details: Artificial intelligence (AI) and machine learning (ML) systems have
advanced significantly in recent years. Despite a wide range of impressive results,
current AI is not intelligent in the biological sense. These systems are limited to
performing only those tasks for which they have been specifically programmed and
trained, and are inherently subject to safety hazards when encountering situations
outside them. The issue is further limiting to DoD applications, where situations
can be unpredictable and the ability to react quickly and adapt to dynamic circum-
stances is of primary importance. The Lifelong Learning Machines (L2M) program
seeks to achieve paradigm-changing developments in AI architectures and ML tech-
niques. The program seeks to develop systems that can learn continuously during
execution and become increasingly expert while performing tasks, are subject to
safety limits, and apply previous skills and knowledge to new situations - without
forgetting previous learning. L2M consists of two technical areas. The first concen-
trates on the development of complete systems and their components; the second
brings together researchers with diverse expertise to explore biological mechanisms
that underlie learning, which will be translated into a new generation of computa-
tional architectures, mechanisms, and algorithms. Discoveries in both technical areas
are expected to generate new methodologies that will allow AI systems to learn and
improve during tasks, apply previous skills and knowledge to new situations, incor-
porate innate system limits, and enhance safety in automated assignments.

124
Computers and Humans Exploring
Software Security (CHESS)
Mr. William Martin

https://www.darpa.mil/program/computers-and-humans-exploring-software-security.html

GPT-3 Summary: The goal of the CHESS program is to develop capabilities to

discover and address vulnerabilities of all types in a scalable, timely, and consistent
manner. This will be done by researching the effectiveness of enabling computers
and humans to collaboratively reason over software artifacts, such as source code and
compiled binaries. If successful, this program will provide a basis for developing new
forms of communication and information sharing between computers and humans,
and enable the discovery of 0-day vulnerabilities at a scale and speed appropriate for
the complex software ecosystem.

Full Details: The Department of Defense (DoD) maintains information systems

that depend on Commercial off-the-shelf (COTS) software, Government off-the-shelf
(GOTS) software, and Free and open source (FOSS) software. Securing this diverse
technology base requires highly skilled hackers who reason about the functionality
of software and identify novel vulnerabilities. This process requires hundreds, if not
thousands of hours of manual effort per discovered vulnerability and does not scale
sufficiently to secure the continuously growing technology base. Hackers use program
analysis techniques and tools to identify and mitigate vulnerabilities, but this pro-
cess requires considerable expertise, manual effort, and time. Automated program
analysis capabilities can reason over only a few vulnerability classes without human
involvement, such as memory corruption or integer overflow, but cannot address the
majority of vulnerabilities. These unaddressed vulnerability types depend on subtle
semantic and contextual information, which is beyond the grasp of modern automa-
tion. Scaling up existing approaches to address the size and complexity of modern
software packages is not possible given the limited number of expert hackers in the
world, much less the DoD.The Computers and Humans Exploring Software Security
(CHESS) program aims to develop capabilities to discover and address vulnerabili-
ties of all types in a scalable, timely, and consistent manner. Achieving the necessary
scale and timelines in vulnerability discovery will require innovative combinations of
automated program analysis techniques with support for advanced computer-human
collaboration. Due to the cost and scarcity of expert hackers, such capabilities must
be able to collaborate with humans of varying skill levels, even those with no pre-
vious hacking experience or relevant domain knowledge.The CHESS program will

125
research the effectiveness of enabling computers and humans to collaboratively rea-
son over software artifacts, such as source code and compiled binaries, with the goal
of finding 0-day vulnerabilities at a scale and speed appropriate for the complex soft-
ware ecosystem upon which the U.S. Government, military, and economy depend.
Achieving these goals will require research breakthroughs in:Developing instrumen-
tation to capture and analyze the process by which hackers reason over software
artifacts to provide a basis for developing new forms of highly effective communica-
tion and information sharing between computers and humans; Creating techniques for
addressing classes of vulnerability that are currently hampered by information gaps
and require human insight and/or contextually sensitive reasoning;Generating repre-
sentations of the information gaps for human collaborators of varying skill levels to
reason over;Integrating human-generated insights into the vulnerability discovery pro-
cess;Emitting a Proof of Vulnerability to confirm existence of the 0-day vulnerability,
and generating a non-disruptive, specific patch to neutralize the 0-day vulnerability;
andSynthesizing vulnerable Challenge Set corpora representative of large, real world,
complex software packages.The CHESS Broad Agency Announcement is available at
https://beta.sam.gov/opp/557cfe6440e774224a008f6923e526f3/view

126
Demonstration Rocket for Agile Cis-
lunar Operations (DRACO)
Maj. Nathan Greiner, USAFDr. Tabitha Dodson

https://www.darpa.mil/program/demonstration-rocket-for-agile-cislunar-operations.html

GPT-3 Summary: GOAL: The goal of the Demonstration Rocket for Agile Cislu-
nar Operations (DRACO) program is to demonstrate a nuclear thermal propulsion
(NTP) system on orbit.

METHODS: The DRACO program will include the baseline design of a NTP
reactor and culminate in a baseline design review. Track B will include development
of an operational system concept to meet operational mission objectives and a
demonstration system design that is traceable to the operational system but focuses
on demonstrating the propulsion subsystem.

IMPACT: NTP offers a high thrust-to-weight ratio around 10,

Full Details: The space domain is essential to modern commerce, scientific discov-
ery, and national defense. Maintaining space domain awareness in cislunar space –
the volume of space between the Earth and the Moon – will require a leap-ahead
in propulsion technology. The goal of the Demonstration Rocket for Agile Cislu-
nar Operations (DRACO) program is to demonstrate a nuclear thermal propulsion
(NTP) system on orbit. NTP uses a nuclear reactor to heat propellant to extreme
temperatures before expelling the hot propellant through a nozzle to produce thrust.
Compared to conventional space propulsion technologies, NTP offers a high thrust-
to-weight ratio around 10,000 times greater than electric propulsion and two-to-five
times greater specific impulse (i.e. propellant efficiency) than chemical propulsion.
The DRACO program anticipates two tracks. Track A will include the baseline design
of a NTP reactor and culminate in a baseline design review. Track B will include
development of an operational system concept to meet operational mission objec-
tives and a demonstration system design that is traceable to the operational system
but focuses on demonstrating the propulsion subsystem. Track B is anticipated to
culminate in a technology maturation plan review for the demonstration system.

127
Smart Non-invasive Assays of Phys-
iology (SNAP)
Dr. Gopal Sarma

https://www.darpa.mil/program/smart-non-invasive-assays-of-physiology.html

GPT-3 Summary: The SNAP program’s goal is to develop a non-invasive de-

vice that can detect molecular biomarkers to predict warfighter physiological states
associated with physical and cognitive readiness. The device will use multiplexed,
multi-omic biomarker detection to predict human performance readiness states in the
context of real-world, DoD-relevant tasks and display an output to the user. The
device developed under the SNAP program would be readily configurable to diverse
DoD needs for readiness assessment, training, and mission planning, thereby provid-
ing an individualized view into real-time warfighter performance state.

Full Details: The Smart Non-invasive Assays of Physiology (SNAP) program aims to
develop a portable, non-invasive device that can detect and utilize molecular biomark-
ers to predict warfighter physiological states associated with physical and cognitive
readiness. The device will leverage multiplexed, multi-omic biomarker detection to
predict human performance readiness states in the context of real-world, DoD-relevant
tasks and display an output to the user. The device developed under the Smart
Non-invasive Assays of Physiology (SNAP) program would be readily configurable to
diverse DoD needs for readiness assessment, training, and mission planning, thereby
providing an individualized view into real-time warfighter performance state.

128
Series Hybrid Electric Propulsion AiR-
craft Demonstration (SHEPARD)
Mr. Steve Komadina

https://www.darpa.mil/program/series-hybrid-electric-propulsion-aircraft-demonstration.html

GPT-3 Summary: The goal of the SHEPARD program is to quickly mature a new
mission-focused aircraft design that can be fielded with the objective of first flight
in 20 months. The SHEPARD vehicle will be larger scale than the GHO X-Plane
(XRQ-72) with operationally representative fuel fraction and mission systems, while
staying below the Group 3 UAS weight limit.

Full Details: The Series Hybrid Electric Propulsion AiRcraft Demonstration (SHEP-
ARD) program is an existing option to the Air Force Research Laboratory (AFRL)
Great Horned Owl (GHO) contract. The SHEPARD program is based on a DARPA
X-Prime construct that takes emerging technologies and burns down system-level inte-
gration risks to provide a minimum viable product to meet an urgent operational need.
The SHEPARD X-Prime program will leverage the GHO hybrid electric architecture
and some of its component technologies, and quickly mature a new mission-focused
aircraft design that can be fielded with the objective of first flight in 20 months. The
SHEPARD vehicle will be larger scale than the GHO X-Plane (XRQ-72) with op-
erationally representative fuel fraction and mission systems, while staying below the
Group 3 UAS weight limit.

129
Modeling Adversarial Activity (MAA)
Dr. William Corvey

https://www.darpa.mil/program/modeling-adversarial-activity.html

GPT-3 Summary: GOAL: Develop mathematical and computational techniques

for modeling adversarial activity in order to produce high-confidence indications and
warnings of efforts to acquire, fabricate, proliferate, and/or deploy weapons of mass
terror (WMTs).
METHODS: MAA will develop the means to create synthetic transaction data sets
that are both realistic and fully releasable to the scientific community, i.e., data that
contains neither personally identifiable information nor restrictions with respect to
classification. Because transaction data may very naturally be modelled using graphs,
mathematical and computational methods to enable large-scale graph analytics in-
cluding graph alignment and merging

Full Details: The goal of the Modeling Adversarial Activity (MAA) program is to
develop mathematical and computational techniques for modeling adversarial activity
for the purpose of producing high-confidence indications and warnings of efforts to
acquire, fabricate, proliferate, and/or deploy weapons of mass terror (WMTs). MAA
assumes that an adversary’s WMT activities will result in observable transactions.
While the probability that any one source alone will reveal a WMT threat may be
low, the probability of detecting a WMT threat can be increased by appropriately
integrating multiple sources of transaction data.MAA requires synthetic transaction
data to drive the development of techniques and tools in ways that will avoid the pri-
vacy and classification issues that can be associated with real-world data. MAA will
develop the means to create synthetic transaction data sets that are both realistic and
fully releasable to the scientific community, i.e., data that contains neither person-
ally identifiable information nor restrictions with respect to classification. Because
transaction data may very naturally be modelled using graphs, mathematical and
computational methods to enable large-scale graph analytics including graph align-
ment and merging, sub-graph detection, and sub-graph matching are of particular
interest.Additional details are available at DARPA-BAA-16-61: Modeling Adversar-
ial Activity.

130
Configuration Security (ConSec)
Mr. John-Francis Mergen

https://www.darpa.mil/program/configuration-security.html

GPT-3 Summary: The Configuration Security (ConSec) program seeks to develop

a system to automatically generate, deploy, and enforce configurations of components
and subsystems for use in military platforms. These configurations should address
system vulnerabilities and minimize attack surfaces while maintaining expected
functionality and performance. To achieve these goals, ConSec will support research
in the following areas:

-Deriving a functional specification for a component and analyzing how set-

tings in its configuration space could impact its functionality, producing useful
configuration semantic models without exhaustive exploration of the configuration
space, and reasoning effectively with incomplete information.
-Constructing models of intended functionality for the composed system with
minimal

Full Details: The growth of the internet-of-things (IoT) and network-connected

composed systems (e.g., aircraft, critical-infrastructure, etc.) has led to unprece-
dented technical diversity in deployed systems. From consumer IoT devices developed
with minimal built-in security, which are often co-opted by malware to launch large
distributed denial of service (DDoS) attacks on internet infrastructure, to remote at-
tacks on Industrial Control System (ICS) devices, these newly connected, composed
systems provide a vast attack surface. While the diversity of functionality and the
scope of what can now be connected, monitored, and controlled over the Internet has
increased dramatically, economies of scale have decreased platform diversity. Inex-
pensive commodity off-the-shelf (COTS) devices have largely replaced single-purpose,
custom devices. The ConSec program seeks to develop a system to automatically
generate, deploy, and enforce configurations of components and subsystems for use in
military platforms. These configurations should address system vulnerabilities and
minimize attack surfaces while maintaining expected functionality and performance.
By viewing each individual component’s configuration as elements of the composed
system’s behavior and security, more secure configurations can be developed and
deployed to enhance security without requiring new software development or large
hardware changes. To achieve these goals, ConSec will support research in the fol-
lowing areas: Deriving a functional specification for a component and analyzing how
settings in its configuration space could impact its functionality, producing useful

131
configuration semantic models without exhaustive exploration of the configuration
space, and reasoning effectively with incomplete information. Constructing models
of intended functionality for the composed system with minimal human-in-the-loop
time by understanding the operational context(s) of the composed system.Ingesting
standard operating procedures (e.g., pilot check-lists) that describe the operator’s
interactions with composed systems and mapping them into functional models of
system behavior.Characterizing attack surfaces stemming from poorly configured or
composed components, and developing approaches to remedy those weaknesses via
configurations.Deploying secure configurations, monitoring them for changes during
operation, and producing context-relevant responses in the event of an identified
change.Designing authoritative and auditable configuration repositories that provide
strong integrity protections.ConSec: BAAConSec: FAQConSec: CUI guideConSec
Proposers Day: SlidesConSec Proposers Day: Overview of Program by Mr. Ja-
cob Torrey (Video)ConSec Proposers Day: QandA on Broad Agency Announcement
(Video)

132
ADvanced Acclimation and Protec-
tion Tool for Environmental Readi-
ness (ADAPTER)
Dr. Paul Sheehan

https://www.darpa.mil/program/advanced-acclimation-and-protection-tool-for-environmental-
readiness.html

GPT-3 Summary: GOAL: To develop systems that give warfighters greater control
over their physiology in order to improve performance.
TECHNOLOGY METHODS: Integrating engineered cells and biochemicals into an
internal, bioelectronic carrier.
IMPACT: Allowing warfighters to signal the carrier to produce and release therapies
as needed to improve performance.

Full Details: Warfighters are travelers, and the bodily inconveniences suffered by
travelers, such as jet lag and traveler’s diarrhea, can seriously degrade operational
readiness and even determine mission success or failure. To maximize warfighter per-
formance, the ADvanced Acclimation and Protection Tool for Environmental Readi-
ness (ADAPTER) program will develop systems that provide warfighters greater con-
trol over their own physiology. This program will integrate engineered cells and bio-
chemicals into an internal, bioelectronic carrier that the warfighter can signal, as
needed, to produce and release therapies that either eliminate the principal cause of
traveler’s diarrhea, pathogenic bacteria, or regulate circadian rhythms disrupted by
mission requirements or jetlag.

133
Hypersonic Air-breathing Weapon
Concept (HAWC)
Mr. Andrew Knoedler

https://www.darpa.mil/program/hypersonic-air-breathing-weapon-concept.html

GPT-3 Summary: The HAWC program seeks to develop and demonstrate critical
technologies for an air-launched hypersonic cruise missile. The program intends to
emphasize efficient, rapid and affordable flight tests to validate key technologies in
three areas: air vehicle feasibility, effectiveness, and affordability.

Full Details: Systems that operate at hypersonic speeds—five times the speed of
sound (Mach 5) and beyond—offer the potential for military operations from longer
ranges with shorter response times and enhanced effectiveness compared to current
military systems. Such systems could provide significant payoff for future U.S. offen-
sive strike operations, particularly as adversaries’ capabilities advance. The Hyper-
sonic Air-breathing Weapon Concept (HAWC) program is a joint DARPA/U.S. Air
Force (USAF) effort that seeks to develop and demonstrate critical technologies to
enable an effective and affordable air-launched hypersonic cruise missile. The pro-
gram intends to emphasize efficient, rapid and affordable flight tests to validate key
technologies. HAWC plans to pursue flight demonstrations to address three critical
technology challenge areas or program pillars—air vehicle feasibility, effectiveness,
and affordability. Technologies of interest include: Advanced air vehicle configura-
tions capable of efficient hypersonic flight Hydrocarbon scramjet-powered propulsion
to enable sustained hypersonic cruise Approaches to managing the thermal stresses
of high-temperature cruise Affordable system designs and manufacturing approaches

134
Modular Optical Aperture Building
Blocks (MOABB)
Dr. Gordon Keeler

https://www.darpa.mil/program/modular-optical-aperture-building-blocks.html

GPT-3 Summary: GOAL: The MOABB program seeks to design all free-space
optics components into a single integrated device.
TECHNOLOGY METHODS: The program calls for devices that are modular and
scalable to apertures of up to 10 centimeters in diameter and that can run at 100
watts of optical power.
IMPACT: Successful completion of the MOABB program could enable rapid 3-D
scanning using devices smaller than a cell-phone camera, high-speed laser commu-
nications without mechanical steering, foliage-penetrating perimeter sensing, remote
wind sensing, and long-range 3-D mapping

Full Details: Free-space optics today requires a telescope, bulk lasers with mechani-
cal beam-steering, detectors, and electronics. The Modular Optical Aperture Building
Blocks (MOABB) program seeks to design all of these components into a single inte-
grated device. In what would be deemed as the most complex electronic-photonic cir-
cuit ever fashioned, the program’s performers will work to create a wafer-scale system
that is 100x smaller and lighter than conventional systems and can steer the optical
beam 1,000x faster than mechanical components. A primary goal of the MOABB
program is the demonstration of integrated electronic-photonic unit cells that can be
tiled together to form large-scale planar apertures. The program calls for devices that
are modular and scalable to apertures of up to 10 centimeters in diameter and that
can run at 100 watts of optical power. Another primary objective is the operation of a
fully-functional, chip-scale LIDAR system capable of three-dimensional (3-D) imaging
at a range of 100 meters. Successful completion of the MOABB program could enable
rapid 3-D scanning using devices smaller than a cell-phone camera, high-speed laser
communications without mechanical steering, foliage-penetrating perimeter sensing,
remote wind sensing, and long-range 3-D mapping. Such capabilities could, in turn,
find applications in intelligence, surveillance, and reconnaissance (ISR), autonomous
navigation, optical communication systems, and robotics.

135
Morphogenic Interfaces (MINT)
Dr. Vishnu Sundaresan

https://www.darpa.mil/program/morphogenic-interfaces.html

GPT-3 Summary: GOAL: The MINT program will investigate innovative ap-
proaches to model and create interfaces that can self-regulate their morphology and
charge transport function.
TECHNOLOGY METHODS: Developing solid/solid charge transfer interface
materials and alloys or alloy-coatings that can minimize transport at the solid/liquid
and solid/vapor interface.
IMPACT: Creating high specific energy solid-state batteries with long
charge/recharge cycles and alloys or alloy-coatings that can exhibit higher fa-
tigue strength in aggressive corrosive environments.

Full Details: The Morphogenic Interfaces (MINT) program aims to enable the de-
velopment of persistent high-performance electrochemical systems by addressing the
irreversible morphological degradation that occurs at the functional material’s in-
terface. The program will investigate innovative approaches to model solid/solid,
solid/liquid, and solid/vapor electrochemical interfaces and apply this knowledge to
create novel interface materials that can self-regulate their morphology and main-
tain desired charge transport function at the electrochemical interface. The MINT
program has two application-centric focus areas. The first focus area will develop
solid/solid charge transfer interface materials and lead to high specific energy solid-
state batteries with long charge/recharge cycles. The second focus area will de-
velop novel alloys or alloy-coatings that can minimize transport (self-passivate) at the
solid/liquid and solid/vapor interface and therefore exhibit higher fatigue strength in
aggressive corrosive environments.

136
Millimeter Wave GaN Maturation
(MGM)
Dr. Timothy Hancock

https://www.darpa.mil/program/millimeter-wave-gan-maturation.html

GPT-3 Summary: GOAL: The Millimeter Wave GaN Maturation (MGM)

program seeks to capitalize on past DARPA investments, maturing the devices
developed under the NEXT program to improve device yield, reduce process cycle
time, and demonstrate key millimeter wave power amplifiers and mixed-signal
circuits through the use of multi-project wafer runs (MPWs).

METHODS: To accomplish its goals, MGM is making strategic yield and re-
producibility improvements that leverage Standard Evaluation Circuits (SEC),
Process Control Monitors (PCM), and short loop fabrication to identify targeted
process step improvements. Additionally, reliability

Full Details: The DARPA Nitride Electronics NeXt-Generation Technology (NEXT)

program had the goal of developing a revolutionary GaN transistor technology to si-
multaneously achieve high-speed and large voltage swing. The expected outcome
of the program was that highly-scaled GaN devices would ultimately benefit the
design of efficient millimeter wave power amplifiers and mixed-signal circuits. The
NEXT program was successful in developing GaN transistors with deeply scaled gate
lengths along with modifications to the heteroepitaxial channel material, ohmic con-
tacts, and overall gate structure, resulting in record performance for GaN devices.
The Millimeter Wave GaN Maturation (MGM) program seeks to capitalize on past
DARPA investments, maturing the devices developed under the NEXT program to
improve device yield, reduce process cycle time, and demonstrate key millimeter wave
power amplifiers and mixed-signal circuits through the use of multi-project wafer runs
(MPWs). To accomplish its goals, MGM is making strategic yield and reproducibil-
ity improvements that leverage Standard Evaluation Circuits (SEC), Process Control
Monitors (PCM), and short loop fabrication to identify targeted process step im-
provements. Additionally, reliability assessment through the use of accelerated life
testing is used to gauge success and Manufacturing Readiness Level (MRL) is evalu-
ated throughout the program. MPWs are used to engage with a diverse set of users
to provide regular feedback on circuit performance, model correlation and design kit
usability.

137
Open, Programmable, Secure 5G (OPS-
5G)
Mr. Tejas Patel

https://www.darpa.mil/program/open-programmable-secure-5g.html

GPT-3 Summary: The goal of the OPS-5G program is to create an open source,
standards-compliant network stack for 5G mobile that is secure by design. The pro-
gram seeks to enable a “plug-and-play” approach to various software components
which reduces reliance on untrusted technology sources. The program also aims to
address the security challenges posed by the introduction of programmability into 5G
networks.

Full Details: 5G is the latest in a series of evolutions in public mobile networking,

with widespread coverage and access on a subscription basis. 5G networks are charac-
terized by improved capabilities across a variety of measures, including throughputs,
latencies, numbers of devices, and battery life. 5G is used to attach small special
purpose devices comprising the Internet of Things (IoT) to the Internet, and the
important and growing number of services provided by the World Wide Web. IoT
devices are often sensors, and 5G access to their data is envisioned to play impor-
tant roles in medicine, manufacturing, and smart cities.Standards processes are used
to maintain interoperability required for a public network, and while many of the
components and component behaviors of 5G have changed little from predecessors,
such as 4G and LTE, the standards for the most futuristic 5G features are those
most in flux. These futuristic features, ranging from network slicing to support for
programmable networks, also present the greatest risk to U.S. national security, as
networks are simultaneously critical infrastructure and the means used for cyberespi-
onage and cyberwarfare.DARPA’s Open, Programmable, Secure 5G (OPS-5G) aims
to address this risk by pursuing research leading to the development of a portable
standards-compliant network stack for 5G mobile that is open source and secure by
design. OPS-5G seeks to create open source software and systems that enable secure
5G and subsequent mobile networks such as 6G. The signature security advantage of
open source software is increased code visibility, meaning that code can be examined,
analyzed, and audited, either manually or with automated tools. In addition, the
portability of open source serves, as a desired side-effect, to decouple the hardware
and software ecosystems. This significantly raises the difficulty of a supply-chain at-
tack and eases the introduction of innovative hardware into the market. The program
seeks to enable a “plug-and-play” approach to various software components which re-

138
duces reliance on untrusted technology sources.Programmable networks bring new
flexibility to 5G, but also introduce novel security challenges. To achieve 5G’s po-
tential benefits, programmability must be implemented and managed carefully. Such
benefits include bespoke networks that are tuned to application needs, as well as
increased network adaptation capabilities. Programmability must also be developed
in ways that avoid rampant opportunities for misuse. Ideally, the introduction of
programmability for 5G will incorporate lessons learned from the well-intentioned in-
troduction of programmability into web browsers, a capability that quickly became
weaponized by malicious actors. OPS-5G aims to improve overall 5G security by in-
creasing trust at a set of soft points that include unmanaged, unattended, long-lived,
and possibly long-forgotten IoT devices. Additionally, OPS-5G aims to address un-
intended and unwanted interactions between network slices and threats from the vast
increases in network scale.More information is available in the OPS-5G BAA.

139
Glide Breaker
Maj. Nathan Greiner, USAF

https://www.darpa.mil/program/glide-breaker.html

GPT-3 Summary: GOAL: The Glide Breaker program aims to develop and demon-
strate technologies to enable defense against hypersonic systems.
TECHNOLOGY METHODS:
IMPACT:

Full Details: The Glide Breaker program began in 2018 to develop and demonstrate
technologies to enable defense against hypersonic systems.

140
Millimeter Wave Digital Arrays (MI-
DAS)
Dr. Timothy Hancock

https://www.darpa.mil/program/millimeter-wave-digital-arrays.html

GPT-3 Summary: GOAL:

The goal of the Millimeter Wave Digital Arrays (MIDAS) program is to cre-
ate the digital array technology that will enable next-generation DoD millimeter
wave systems.

METHODS:

The program is focused on two key technical areas – one that addresses the
development of core CMOS chips, and a second that seeks to integrate those CMOS
chips with the rest of the aperture.

IMPACT:

Achieving the goals of the program will result in a 16-element tile building
block that can be used to form larger apertures.

Full Details: There is increasing interest in making broader use of the millimeter
wave frequency band for communications on small mobile platforms where narrow
antenna beams from small radiating apertures provide enhanced communication secu-
rity. Today’s millimeter wave systems, however, are not user friendly and are designed
to be platform specific, lacking interoperability and are thus reserved for only the most
complex platforms. To expand the use of millimeter wave phased-arrays and make
them broadly applicable across Department of Defense (DoD) systems, many techni-
cal challenges must be addressed, including wideband frequency coverage, precision
beam pointing, user discover and mesh networking. The goal of the Millimeter-Wave
Digital Arrays (MIDAS) program is to create the digital array technology that will
enable next-generation DoD millimeter wave systems. The program seeks to develop
element-level digital beamforming that will support emerging multi-beam communi-
cations and directional sensing of the electromagnetic environment in the 18-50 GHz
band. To accomplish its goals, MIDAS is focused on two key technical areas – one
that addresses the development of core CMOS chips, and a second that seeks to inte-

141
grate those CMOS chips with the rest of the aperture. The challenge associated with
developing the CMOS chips is achieving the very low power consumption needed, but
with a large amount of RF, mixed-signal, and digital functionality integrated into
a very small space. The second challenge is to implement the low-noise amplifiers,
power amplifiers, transmit/receive switches, and antenna all with high efficiency and
compact chip-scale assembly techniques. Achieving both of these goals will result in
a 16-element tile building block that can be used to form larger apertures. In the
third phase of the program, the focus will be on the creation of a 256-element array
that is composed of the MIDAS tile building blocks and is capable of demonstrating
digital multi-beam functionality over the full band of interest.

142
Three Dimensional Monolithic System-
on-a-Chip (3DSoC)
Dr. Jason Woo

https://www.darpa.mil/program/three-dimensional-monolithic-system-on-a-chip.html

GPT-3 Summary: GOAL: The 3DSoC program seeks to develop the monolithic
3D technology required to build logic, memory, and input/output (I/O) on a single
die using a legacy lithography node while improving performance by more than 50X
when compared with leading edge technology nodes.

TECHNOLOGY METHODS: 3DSoC seeks to develop fabrication technology

as well as the design flows required to take advantage of the technology’s capabilities.

IMPACT: The resulting fabrication technology is intended to be based at a

U.S. manufacturing facility, which will allow

Full Details: Deployed electronic systems increasingly require advanced process-

ing capabilities, however the time and power required to access system memory –
commonly referred to as the “memory bottleneck” – takes a significant toll on their
performance. Any substantial improvement in electronic system performance will re-
quire a radical reduction in memory access time and overall dynamic power of the
system. The use of a monolithic three-dimensional system-on-chip (SoC) stack to in-
tegrate memory and logic is one approach that could dramatically alter the memory
bottleneck challenge. However, developing an effective 3D SoC stack would require
increasing the width of buses to memory through much finer pitch interconnects, while
simultaneously decreasing the resistor-capacitor (RC) delays through much shorter
interconnect lines in the SoC To address the memory bottleneck problem, the Three
Dimensional Monolithic System-on-a-Chip (3DSoC) program seeks to develop the
monolithic 3D technology required to build logic, memory, and input/output (I/O)
on a single die using a legacy lithography node while improving performance by more
than 50X when compared with leading edge technology nodes. To achieve its goals,
3DSoC seeks to develop fabrication technology as well as the design flows required to
take advantage of the technology’s capabilities. The 3DSoC program aims to drive
the future of computation architecture, removing the limitations of the current 2D
approach to SoC design and manufacturing. The resulting fabrication technology is
intended to be based at a U.S. manufacturing facility, which will allow Department
of Defense (DoD) entities to have easy access to the technology, tools, and manufac-

143
turing capabilities to explore and build future computation systems that will deliver
50X better performance at power results and improve the performance of future DoD
systems.

144
Context Reasoning for Autonomous
Teaming (CREATE)
Mr. Aaron Kofford

https://www.darpa.mil/program/context-reasoning-for-autonomous-teaming.html

GPT-3 Summary: GOAL: The CREATE program will allow autonomous machines
to work together in teams to better complete unplanned missions.
TECHNOLOGY METHODS: The program will explore the use of artificial intel-
ligence to enable a system of heterogeneous, contextually-aware agents to act in a
decentralized manner.
IMPACT: The program will improve the ability of autonomous machines to complete
unplanned missions.

Full Details: CREATE aims to explore the utility of artificial intelligence (AI) on
the autonomous formation of scalable machine-to-machine teams capable of reacting
to and learning from unexpected missions in the absence of centralized communication
and control. CREATE seeks to develop the theoretical foundations of autonomous
AI teaming to enable a system of heterogeneous, contextually-aware agents to act
in a decentralized manner and satisfy multiple, simultaneous and unplanned mis-
sions goals. Agents within the team will have mechanism for regulation to ensure
(favorable) emergent behavior of the team to: (1) better ensure the desired mission
outcome; and (2) bound the cost of unintended adverse action or regret. CREATE
aims to explore the function and utility of encoding common knowledge (e.g. gen-
eral places and things), procedural knowledge (e.g. user’s manuals and playbooks),
and learned knowledge (evolving experiential, or semantically inferred information).
Additionally, the program will explore the accuracy and value of decentralized ma-
chines decisions made form local observations and any available global context with
an emphasis on decisions related to unplanned missions. To mitigate potential un-
desirable action while maintaining an interesting level of autonomy, the program will
investigate the correct balance of hard coded and contextual safeguards.

145
Advanced Full Range Engine (AFRE)
Maj. Nathan Greiner, USAF

https://www.darpa.mil/program/advanced-full-range-engine.html

GPT-3 Summary: The goal of the AFRE program is to develop a new aircraft
propulsion system that can operate over the full range of speeds required for low-
speed takeoff and hypersonic flight. The program will initially conduct system design,
sub- and large-scale component development, and ground demonstration, and will
culminate with a large-scale, integrated test series or a propulsion wind tunnel free-
jet test of the integrated low- and high-speed flowpaths.

Full Details: In the decades-long quest to develop reusable aircraft that can reach
hypersonic speeds – Mach 5 (approximately 3,300 miles per hour/5,300 kilometers
per hour) and above – engineers have grappled with two intertwined, seemingly in-
tractable challenges: The top speed of traditional jet-turbine engines maxes out at
roughly Mach 2.5, while hypersonic engines such as scramjets cannot provide effective
thrust at speeds much below Mach 3.5. This gap in capability means that any air-
breathing hypersonic vehicles developed today would use disposable rockets for one-
time boosts up to operating speed, limiting the vehicles’ usefulness. To help remove
these constraints and lay the framework for routine hypersonic flight with reusable
vehicles, DARPA has launched its Advanced Full Range Engine (AFRE) program.
AFRE seeks to develop and demonstrate a new aircraft propulsion system that could
operate over the full range of speeds required from low-speed takeoff through hyper-
sonic flight. AFRE aims to explore a turbine-based combined cycle (TBCC) engine
concept, which would use a turbine engine for low-speed operations and a dual-mode
ramjet – which would work efficiently whether the air flowing through it is subsonic
(as in a ramjet) or supersonic (as in a scramjet) – for high-speed operations. The
AFRE program will initially conduct system design, sub- and large-scale component
development, and ground demonstration, and will culminate with a large-scale, inte-
grated test series or a propulsion wind tunnel free-jet test of the integrated low- and
high-speed flowpaths.

146
Neural Evidence Aggregation Tool
(NEAT)
Dr. Bartlett Russell

https://www.darpa.mil/program/neural-evidence-aggregation-tool.html

GPT-3 Summary: The NEAT program will develop a new cognitive science tool
that can identify people at risk of suicide by using preconscious brain signals. This will
help clinicians to assess patients more accurately and provide earlier interventions.

Full Details: Since Sept. 11, 2001, more than 30,000 active duty members and
veterans have taken their own lives – four times as many as those killed in post-
911 military operations.1 Effective behavioral health assessment is a mission-critical
capability requiring novel tools to identify and help those at risk. Current methods to
detect early signs of behavioral and mental health risk factors rely on self-reporting
and screening questionnaires, which cannot reliably predict suicidality.The Neural
Evidence Aggregation Tool (NEAT) program aims to overcome current limitations
by developing a new cognitive science tool that identifies people at risk of suicide
by using preconscious brain signals rather than asking questions and waiting for
consciously filtered responses. By aggregating preconscious brain signals to stimuli,
NEAT would determine what a person believes to be true, false, or indeterminate
about specific types of knowledge that could be used to detect signs of depression,
anxiety, or suicidal ideation earlier and more reliably than ever before. If successful,
NEAT will not only significantly augment behavioral health screening, but it could
also serve as a new way to assess ultimate treatment efficacy, since patients will often
tell their clinicians what they think the clinician wants to hear rather than how they
are truly feeling. Ultimately, NEAT intends to augment current behavioral health
screening programs by providing clinicians with previously unavailable information
to enable earlier interventions and more reliable measures of successful treatment.
1 High Suicide Rates among United States Service Members and Veterans of the
Post9/11 Wars, Thomas Howard Suitt, III, Boston University, June 2021

147
Bridging the Gap Plus (BG+)
CDR Jean-Paul Chretien

https://www.darpa.mil/program/bridging-the-gap-plus.html

GPT-3 Summary: The BG+ program seeks to develop new approaches for
treating spinal cord injury by integrating injury stabilization, regenerative therapy,
and functional restoration. To achieve this combinatorial approach, BG+ teams will
build two systems of implantable and adaptive devices:

The first system will reduce injury effects during the early phases of spinal
cord injury. This system will conduct real-time biomarker monitoring and interven-
tion to stabilize – and where possible, rebuild – the neural communications pathways
at the site of injury.

The second system will primarily address recovery of function in the final phase of
spinal cord injury by deploying networks of stimulation and

Full Details: The Bridging the Gap Plus (BG+) program aims to develop new
approaches for treating spinal cord injury by integrating injury stabilization, regen-
erative therapy, and functional restoration. To achieve this combinatorial approach,
BG+ teams will build two systems of implantable and adaptive devices. The first
system will reduce injury effects during the early phases of spinal cord injury. This
system will consist of active devices that conduct real-time biomarker monitoring
and intervention to stabilize – and where possible, rebuild – the neural communica-
tions pathways at the site of injury. Biomarker monitoring will provide previously
unavailable diagnostic information for automated or clinician-directed interventions.
The second system will primarily address recovery of function in the final phase of
spinal cord injury by deploying networks of stimulation and/or recording devices on
the nervous system or relevant end organs to effectively “bridge the gap” of the spinal
cord injury. The BG+ program seeks to change the paradigm for spinal cord injury
treatment. Successful BG+ devices will mitigate the early effects of injury, enhance
clinical awareness and therapeutic options at the injury site, accelerate natural long-
term recovery processes, and restore multiple nervous system functions.

148
Gamma Ray Inspection Technology
(GRIT)
Dr. Mark Wrobel

https://www.darpa.mil/program/gamma-ray-inspection-technology.html

GPT-3 Summary: GOAL: Develop new approaches to produce high-intensity,

tunable, and narrow-bandwidth gamma rays in a compact form factor.

TECHNOLOGY METHODS: New approaches to gamma ray production.

IMPACT: Help discover smuggled nuclear materials in cargo, provide new

non-destructive inspection techniques at various scales, and enable new medical
diagnostics and therapies.

Full Details: The Gamma Ray Inspection Technology (GRIT) program seeks trans-
formational approaches to achieving high-intensity, tunable, and narrow-bandwidth
gamma ray production, but in a compact form factor suitable for transporting the
source to where the capability is needed. Such sources have the potential to help
discover smuggled nuclear materials in cargo, provide new non-destructive inspection
techniques at various scales, and enable new medical diagnostics and therapies.

149
Biostasis
Dr. Anne Cheever

https://www.darpa.mil/program/biostasis.html

GPT-3 Summary: The Biostasis program’s goal is to develop novel chemical biol-
ogy approaches to reversibly and controllably slow biological systems without cold-
chain to stabilize and protect their functional capacity until medical intervention is
possible. This would extend the time for lifesaving medical treatment, often referred
to as “the Golden Hour,” following traumatic injury or acute infection, thus increasing
survivability for military personnel operating in far-forward conditions with limited
access to medical professionals or trauma centers.

Full Details: The Biostasis program aims to extend the time for lifesaving med-
ical treatment, often referred to as “the Golden Hour,” following traumatic injury
or acute infection, thus increasing survivability for military personnel operating in
far-forward conditions with limited access to medical professionals or trauma centers.
To do so, Biostasis is developing novel chemical biology approaches that reversibly
and controllably slow biological systems without cold-chain to stabilize and protect
their functional capacity until medical intervention is possible. Biostasis is investi-
gating novel applications of polymer chemistry, protein engineering, and deep cell
activity monitoring to alter the time course of pathological processes associated with
tissue damage and infection to delay the onset of irreversible damage. Researchers
are investigating approaches that are not dependent upon reducing temperature and
that scale from preservation of simple biological therapeutics such as antibodies and
enzymes to whole cells and tissues. The program seeks to generate proof-of-concept,
benchtop technologies and experimentally validate them in simple biological systems.
DARPA will work with federal health and regulatory agencies as the program ad-
vances to develop a pathway for potential, future human medical use of successful
Biostasis technologies. By the end of the program, DARPA hopes to have multiple
tools for reducing the risk of permanent damage or death following acute injury or in-
fection. Biostasis technologies could also extend the shelf-life of temperature-sensitive
therapeutics, such as blood products, enzyme preparations, and drugs.

150
Advanced Plant Technologies (APT)
Dr. Kristen Jordan

https://www.darpa.mil/program/advanced-plant-technologies.html

GPT-3 Summary: The goal of the APT program is to develop plants that can act as
sensors for a variety of threats, including chemical, biological, radiological, and nuclear
threats. The program is researching ways to modify plants so that they are more
sensitive to these threats and can send signals to be monitored by existing ground-,
air-, or space-based hardware. If successful, this could provide a more effective and
efficient way to detect and respond to threats.

Full Details: The Advanced Plant Technologies (APT) program seeks to develop
plants capable of serving as next-generation, persistent, ground-based sensor tech-
nologies to protect deployed troops and the homeland by detecting and reporting on
chemical, biological, radiological, nuclear, and explosive (CBRNE) threats. Such bio-
logical sensors would be effectively energy-independent, increasing their potential for
wide distribution, while reducing risks associated with deployment and maintenance
of traditional sensors. These technologies could also potentially support humanitarian
operations by, for example, detecting unexploded ordnance in post-conflict settings.
DARPA’s technical vision for APT is to harness plants’ innate mechanisms for sens-
ing and responding to environmental stimuli, extend that sensitivity to a range of
signals of interest, and engineer discreet response mechanisms that can be remotely
monitored using existing ground-, air-, or space-based hardware. To succeed, APT
must ensure that modified plants are safe, robust, and self-sustaining in their envi-
ronments. The program hinges on the advancement of technologies for performing
multiple, complex modulations to plants, without sacrificing their environmental fit-
ness. Although APT pursues technology for eventual deployment, the initial research
is conducted entirely in contained facilities. If the research is successful, later-phase
field trials would take place under the auspices of the U.S. Department of Agricul-
ture’s Animal and Plant Health Inspection Service following all standard protocols
for plant biosafety.

151
Bioelectronics for Tissue Regenera-
tion (BETR)
Dr. Paul Sheehan

https://www.darpa.mil/program/bioelectronics-for-tissue-regeneration.html

GPT-3 Summary: The Bioelectronics for Tissue Regeneration (BETR) program

will develop technology to speed up warfighter recovery by directly intervening in
wound healing. This will be done by building an adaptive system that uses actuators
to stimulate tissue, sensors to track the body’s response to stimulation, and adaptive
learning algorithms to integrate sensor data and dictate intervention to the actuators.
Ultimately, the goal is to provide medical interventions with the necessary sophisti-
cation to more quickly restore complex human tissues after injury. If the program is
successful, it could have a major impact on the health and wellbeing of warfighters.

Full Details: The Bioelectronics for Tissue Regeneration (BETR) program will
develop technology aimed at speeding warfighter recovery, and thus resilience, by di-
rectly intervening in wound healing. To do this, researchers will build an adaptive
system that uses actuators to biochemically or biophysically stimulate tissue, sen-
sors to track the body’s complex response to that stimulation, and adaptive learning
algorithms to integrate sensor data and dictate intervention to the actuators. After es-
tablishing this closed-loop control over physiological processes, BETR researchers will
integrate these devices into a single platform that guides the tissue in real-time along
an optimal growth pathway. Ultimately, through the BETR program DARPA aims
to provide medical interventions with the necessary sophistication to more quickly
restore complex human tissues after injury. The effort notably focuses on injuries rel-
evant to the warfighter such as blast and burn damage to skin, muscle, and nerves. If
the program succeeds, warfighters could return to duty sooner and more fully healed.

152
Knowledge-directed Artificial Intel-
ligence Reasoning Over Schemas (KAIROS
Dr. William Corvey

https://www.darpa.mil/program/knowledge-directed-artificial-intelligence-reasoning-
over-schemas.html

GPT-3 Summary: GOAL: The KAIROS program seeks to develop a schema-based

AI system that can identify complex events and bring them to the attention of users.

TECHNOLOGY METHODS: The first stage entails the development of auto-

mated approaches for learning schemas from big data, and the second stage focuses
on the development of automated technologies that apply these schemas to multi-
media/multilingual information to discover and extract complex events of interest to
KAIROS users.

IMPACT: The goal of this program is to improve the ability of AI systems to

identify complex events and bring them to

Full Details: Rapid comprehension of world events is essential for informing U.S.
national security - a task that becomes more difficult as the amount of unstructured,
multimedia information grows exponentially. Humans make sense of events by or-
ganizing them into narrative structures that occur frequently. These structures are
abstracted into schemas, which are organized units of knowledge that represent a
pattern of memory used in human cognition. First-wave, or rule-based, symbolic
reasoning, AI systems that incorporated hand-crafted schemas were unable to scale
when matching those schemas to real-world data. Second-wave, or machine learning,
AI systems require far too many manually-produced, annotated examples as training
data for supervised machine learning methods to be practical.The Knowledge-directed
Artificial Intelligence Reasoning Over Schemas (KAIROS) program seeks to develop
a schema-based AI system that can identify complex events and bring them to the at-
tention of users. KAIROS aims to understand complex events described in multimedia
inputs by developing a semi-automated system that identifies, links, and temporally
sequences their subsidiary elements, the participants involved, as well as the com-
plex event type. Events of interest either create changes that have significant impact
on national security or participate in causal chains that produce such impacts.The
KAIROS program seeks to overcome the scaling limitations of prior approaches in
two stages. As shown in the figure below, the first stage entails the development

153
of automated approaches for learning schemas from big data, and the second stage
focuses on the development of automated technologies that apply these schemas to
multi-media/multilingual information to discover and extract complex events of in-
terest to KAIROS users. A KAIROS Proposers Day was held January 9, 2019 in
Arlington, VA.

154
Hunter
CDR Greg Sutton

https://www.darpa.mil/program/hunter.html

GPT-3 Summary: GOAL: The Hunter program seeks to develop an innovative

concept for the delivery of advanced undersea payloads from extra-large unmanned
underwater vehicles (XLUUVs).

TECHNOLOGY METHODS: The Hunter program will be executed in phases.

Phase 1 seeks to design and build the payload delivery device to fit inside a
government-provided payload module. Phases 2 and 3 aim to support integration of
the payload delivery device into the XLUUV and perform testing.

IMPACT: Successful completion of the Hunter program would enable the de-
livery of advanced payloads from XLUUVs, which would

Full Details: The Hunter program seeks to develop an innovative concept for the
delivery of advanced undersea payloads from extra-large unmanned underwater ve-
hicles (XLUUVs). The Hunter program will be executed in phases. Phase 1 seeks
to design and build the payload delivery device to fit inside a government-provided
payload module. Phases 2 and 3 aim to support integration of the payload delivery
device into the XLUUV and perform testing.

155
Biomanufacturing: Survival, Utility,
and Reliability beyond Earth (B-SURE)
Dr. Anne Cheever

https://www.darpa.mil/program/biomanufacturing-survival-utility-and-reliability-beyond-
earth.html

GPT-3 Summary: GOAL: The goal of this program is to improve the resiliency
of the supply chain and enable future orbital missions.

TECHNOLOGY METHODS: This program will assess how well microbial sys-
tems can utilize alternative feedstock and how space conditions impact the
performance of common biomanufacturing host organisms.

IMPACT: The successful completion of this program will provide data that
will be critical in enabling future orbital missions.

Full Details: The DoD has a role in orbital and lunar missions as defined by the U.S.
Space Force (USSF) Space Capstone Publication. To succeed in this role, there is a
critical DoD need for the continued development and expansion of orbital manufactur-
ing to enable and ensure supply chain resiliency, sustained technological superiority,
and asset security and repair for current and future operations. The Biomanufactur-
ing: Survival, Utility, and Reliability beyond Earth (B-SURE) Program will evaluate
basic biological questions related to alternative feedstock utilization by microbial
systems and how variable gravity and galactic cosmic radiation impacts cellular per-
formance of common biomanufacturing host organisms. By addressing these areas,
B-SURE will provide critical information and data on the future of biomanufacturing
beyond Earth.

156
Nucleic acids On-demand Worldwide
(NOW)
Dr. Amy Jenkins

https://www.darpa.mil/program/nucleic-acids-on-demand-worldwide.html

GPT-3 Summary: The goal of the NOW program is to develop a mobile platform
for manufacturing nucleic acid therapeutics (DNA and RNA) on-demand. This will
allow for the rapid production of MCMs in the event of a bio-threat attack or outbreak
of an infectious disease. The mobile platform will be able to operate anywhere the
military is deployed, providing immediate response to threats.

Full Details: The Nucleic acids On-demand Worldwide (NOW) program aims to
develop a mobile medical countermeasure (MCM) manufacturing platform for use in
stabilization and humanitarian operations to rapidly produce, formulate, and package
hundreds of doses of nucleic acid therapeutics (DNA and/or RNA). These on-demand
MCMs can also better prepare our deployed, field-forward forces against bio-threat
attacks and emerging infectious disease. Fielding such a nucleic acid manufacturing
platform will enable deployment of pure, sequence accurate MCMs in days, rather
than months to years. Further, the mobile nature of the NOW platform will provide
immediate threat response anywhere the military operates, mitigate surprise from
disease threats, and ensure a resilient, readily accessible nucleic acid MCM manufac-
turing pipeline.

157
Securing Information for Encrypted
Verification and Evaluation (SIEVE)
Dr. Joshua Baron

https://www.darpa.mil/program/securing-information-for-encrypted-verification-and-
evaluation.html

GPT-3 Summary: GOAL: The SIEVE program seeks to create more efficient zero-
knowledge proofs to enable verification of complex statements without revealing sen-
sitive details.
TECHNOLOGY METHODS: The program will focus on increasing the expressivity
of problem statements for which ZK proofs can be constructed, as well as on increas-
ing the efficiency of ZK proof technology.
IMPACT: The goal is to enable verification of capabilities relevant to the Department
of Defense without revealing the sensitive details associated with those capabilities.

Full Details: A zero-knowledge (ZK) proof is an interactive protocol between a

prover and a verifier. The prover creates a statement that they want the verifier
to accept, using knowledge that will remain hidden from the verifier. Recent re-
search has substantially increased the efficiency of ZK proofs, enabling real-world
use, primarily by cryptocurrencies. While useful for cryptocurrencies, the ZK proofs
created are specialized for this task and do not necessarily scale for transactions that
are more complex. For highly complex proof statements like those that the Depart-
ment of Defense (DoD) may wish to employ, novel and more efficient approaches are
needed.The Securing Information for Encrypted Verification and Evaluation (SIEVE)
program seeks to advance the state of the art in ZK proofs to enable complex, DoD-
relevant applications. SIEVE will use ZK proofs to enable the verification of capa-
bilities relevant to the DoD without revealing the sensitive details associated with
those capabilities. SIEVE will aim to accomplish this goal by dramatically increas-
ing the expressivity of problem statements for which ZK proofs can be constructed.
SIEVE will also focus on increasing the efficiency of ZK proof technology to enable
large, complex proof statements (e.g., billions of gates or more, where the statement
natively consists of probabilistic, indeterminate-branching conditions). SIEVE will
demonstrate the feasibility of encoding complex, DoD-relevant statements into in-
termediate representations (IRs) that can then be used to create efficient ZK proofs
for those statements.Additionally, in order to ensure the relevance of ZK proofs for
the foreseeable future, including the case where a cryptographically-relevant quantum
computer were to exist, SIEVE will focus on substantially decreasing the asymptotic

158
complexity of post-quantum ZK proof techniques, specifically ZK proofs that 1) rely
on post-quantum hardness assumptions for their security and/or 2) reason about
statements of relevance to post-quantum cryptography.

159
Space-Based Adaptive Communica-
tions Node (Space-BACN)
Dr. Greg Kuperman

https://www.darpa.mil/program/space-based-adaptive-communications-node.html

GPT-3 Summary: GOAL: The Space-BACN program will develop a reconfigurable,

multi-protocol intersatellite optical communications terminal.
TECHNOLOGY METHODS: Develop a terminal that is low size, weight, power, and
cost (SWaP-C), easy to integrate, and will have the ability to connect heterogeneous
constellations that operate on different optical intersatellite link (OISL) specifications.
IMPACT: Eliminate stovepipes and “connect space,” which will in turn enable the
joint all-domain fight.

Full Details: The goal of Space-BACN is to create a reconfigurable, multi-protocol

intersatellite optical communications terminal that is low size, weight, power, and
cost (SWaP-C), easy to integrate, and will have the ability to connect heterogeneous
constellations that operate on different optical intersatellite link (OISL) specifications
that otherwise would not be able to communicate. In simpler terms, the goal of this
program is to eliminate stovepipes and “connect space,” which will in turn enable
the joint all-domain fight. Space-BACN aims to overcome the perpetual lack of full
on-orbit interoperability with current and future space communications. Since pro-
liferated space is so new, there is no standardization of communications or OISLs in
this domain. New constellations that are in different stages of acquisition are procur-
ing single-waveform cross-link communication systems that meet their mission’s or
business objective’s particular needs to interconnect their own constellation. These
single-waveform systems are only capable of talking to other systems that support
that particular waveform, almost exclusively consist of custom-made components, and
have little to no reconfigurability. While most waveforms operate within the same
wavelength band, they differ in essentially all implementation details, including spe-
cific wavelengths, polarization, clock rate, spatial acquisition sequence, modulation
format, framing, and error correction coding. With each constellation acquiring its
own single-waveform OISL, the space domain has become severely fragmented with
isolated islands of connectivity. Space-BACN’s multi-standard optical terminal that
can be reconfigured on-orbit will enable communications across different standards,
and connect these otherwise isolated constellations.To learn more about the Space-
BACN program, visit the program webpage here: https://www.darpa.mil/work-with-
us/space-based-adaptive-communications-node

160
AI-assisted Climate Tipping-point Mod-
eling (ACTM)
Dr. Joshua Elliott

https://www.darpa.mil/program/ai-assisted-climate-tipping-point-modeling.html

GPT-3 Summary: GOAL: The goal of this program is to use AI to improve climate
models, specifically to make them more accurate and efficient.
METHODS: This will be done by developing new methods for modeling complex
processes and for assimilating diverse data.
IMPACT: The hope is that this will lead to better predictions of climate change,
including sudden tipping points, and help policy makers make better decisions.

Full Details: The program objective is to explore AI-assisted modeling of complex

processes related to climate. The specific goals of this effort are to: Explore the use
of third wave AI methods to enhance models of complex interconnected processes. In
particular, to develop hybrid AI models of the climate and Earth system that cap-
ture missing physical, chemical, or biological processes with sufficient computational
efficiency to explore decadal scale effects and characterize tipping points and bifurca-
tions.Develop methods to assimilate diverse data into models and estimate the “value
of new data” to enhance confidence in target-specific forecasts relative to state-of-the-
art (SOTA) techniques.Climate change, whether natural or human-driven, has huge
potential impacts on geopolitical and economic stability, food and water security, and
DoD missions and operations. Current climate models of highly complex, underlying
physical processes are computationally intensive and provide limited actionable guid-
ance to policy makers, especially on the risks and causes of sudden tipping-points,
runaway feedback loops, and the strategic implications of potential adversarial activ-
ity. Third wave AI methods (e.g., neuro-symbolic hybrid AI models that can incorpo-
rate context and can extract causal factors and internal structure) have the potential
to improve the accuracy of climate forecasts as well as improve the predictability of
tipping points and to provide actionable guidance on new data when predictability
remains poor. Faster learned models, particularly when used in conjunction with full-
scale physics models for validation, will enable policy makers to better explore the
climate impacts and risks of policy decisions. Quantifying climate risks is essential to
prepare for a range of scenarios such as those related to DoD planning and decision
support (e.g., arctic strategy/defense, regional destabilization, global power/economic
realignment, base/force locations, and extreme weather threats) and to identify new
potential high-value observations (e.g., stratospheric vs. ocean surface vs. deep ocean

161
vs. arctic, etc.) to enhance confidence in forecasts. The 18-month AIE program will
develop hybrid AI methods in Phase 1 (12-months) using open and rich datasets; and
in Phase 2 (6 months), the program will apply the developed methods to quantify
the “value of new data” for target-specific forecasts and establish the benefits of new
methods relative to SOTA.

162
Atomic Clock with Enhanced Sta-
bility (ACES)
Dr. Dev Palmer

https://www.darpa.mil/program/atomic-clock-with-enhanced-stability.html

GPT-3 Summary: GOAL: The ACES program will develop next-generation

chip-scale atomic clocks (CSAC) that are 1000 times more accurate than existing
CSAC technology.

TECHNOLOGY METHODS: The ACES program will explore alternative CSAC

physics architectures and include engineering efforts to demonstrate SWaP reduction
of laboratory-proven atomic clock technologies.

IMPACT: The ACES program could substantively impact future ACES clock
architectures, making them more accurate and reliable.

Full Details: Precise timing is essential across DoD systems, including communica-
tions, navigation, electronic warfare, intelligence systems reconnaissance, and system-
of-systems platform coordination, as well as in national infrastructure applications in
commerce and banking, telecommunications, and power distribution. Improved clock
performance throughout the timing network, particularly at point-of-use, would en-
able advanced collaborative capabilities and provide greater resilience to disruptions
of timing synchronization networks, notably by reducing reliance on satellite-based
global navigation satellite system (GNSS) timing signals. The Atomic Clock with En-
hanced Stability (ACES) program aims to develop next-generation, battery-powered
chip-scale atomic clocks (CSAC) with 1000X improvement in key performance param-
eters compared to existing CSAC technology.First-generation CSACs, developed un-
der prior DARPA programs, are now available commercially and have demonstrated
the feasibility and value of CSAC technology across DoD and civilian applications.
These devices offer unprecedented timing stability within their domain of size, weight,
and power (SWaP), but their performance is fundamentally limited—particularly in
the key performance metrics of temperature sensitivity (tempco), long-term frequency
aging, and turn-on to turn-on reproducibility (retrace)—due to the physics associated
with the design of these devices.The ACES program is exploring alternative CSAC
physics architectures with the objective of demonstrating a CSAC with 1000X im-
provement in tempco, aging, and retrace. The program includes engineering efforts
to demonstrate SWaP reduction of laboratory-proven atomic clock technologies and

163
to deliver ACES prototype clocks. The program also includes basic research efforts to
explore novel component technologies and alternative physics approaches that have
the potential to substantively impact future ACES clock architectures.

164
Intelligent Design of Electronic As-
sets (IDEA)
Dr. James Wilson

https://www.darpa.mil/program/intelligent-design-of-electronic-assets.html

GPT-3 Summary: GOAL: The goal of the IDEA program is to develop a

general purpose hardware compiler that can translate source code or schematics into
physical layout (GDSII) in less than 24 hours, without the need for large design teams.

TECHNOLOGY METHODS: The IDEA program will leverage machine learn-

ing, optimization algorithms, and expert systems to create the compiler.

IMPACT: This program has the potential to reduce the cost and complexity
barriers associated with leading-edge electronic design.

Full Details: Next-generation intelligent systems supporting Department of Defense

(DoD) applications like artificial intelligence, autonomous vehicles, shared spectrum
communication, electronic warfare, and radar require processing efficiency that is or-
ders of magnitude beyond what is available through current commercial electronics.
Reaching the performance levels required by these DoD applications however will
require developing highly complex system-on-chip (SoC) platforms that leverage the
most advanced integrated circuit technologies. Despite advances in electronic design
automation (EDA) tools, the complexity associated with the design and verification
of integrated circuits (ICs) continues to increase rapidly due in part to the steady
progress of Moore’s Law. To help meet design requirements, commercial electronics
manufacturers creating advanced hardware solutions will employ large teams of de-
signers, each with expertise in a specific facet of the design flow. For the DoD however,
researchers and development teams do not have the resources available to effectively
execute such a strategy, resulting in hardware design cycles that are two to three
times longer than commercial establishments.To overcome the design expertise gap
and keep pace with the exponential increase in chip complexity, the Intelligent Design
of Electronic Assets (IDEA) program seeks to develop a general purpose hardware
compiler for no-human-in-the-loop translation of source code or schematic to physi-
cal layout (GDSII) for SoCs, System-In-Packages (SIPs), and Printed Circuit Boards
(PCBs) in less than 24 hours. The program aims to leverage advances in applied
machine learning, optimization algorithms, and expert systems to create a compiler
that could allow users with no prior design expertise to complete physical design at

165
the most advanced technology nodes. The goal of the IDEA program is to provide the
DoD with a path to rapid development of next-generation electronic systems without
the need for large design teams, reducing the cost and complexity barriers associated
with leading-edge electronic design.

166
Artificial Social Intelligence for Suc-
cessful Teams (ASIST)
Dr. Joshua Elliott

https://www.darpa.mil/program/artificial-social-intelligence-for-successful-teams.html

GPT-3 Summary: GOAL: The ASIST program seeks to develop foundational AI

theory and systems that demonstrate the basic machine social skills needed to infer
the goals and situational knowledge of human partners, predict what they will need,
and offer context-aware actions in order to perform as adaptable and resilient AI
teammates.
TECHNOLOGY METHODS: The ASIST program will create agents that demon-
strate a machine theory of mind and the ability to participate in an effective team by
representing and helping to maintain shared models.
IMPACT: The ASIST program will create a testbed for evaluating these agents using
customizable open-world environments and

Full Details: Humans intuitively combine pre-existing knowledge with observations

and contextual clues to construct rich mental models of the world around them and
use these models to evaluate goals, perform thought experiments, make predictions,
and update their situational understanding. When the environment contains other
people, humans use a skill called theory of mind (ToM) to infer their mental states
from observed actions and context, and predict future actions from those inferred
states. When humans form teams, these models can become extremely complex.
High-performing teams naturally align key aspects of their models to create shared
mental models of their environment, equipment, team, and strategies. ToM and the
ability to create shared mental models are key elements of human social intelligence.
Together, these two skills form the basis for human collaboration at all scales, whether
the setting is a playing field or a military mission. Artificial intelligence (AI) tech-
nologies have made little progress in understanding the most important component of
the environments in which they operate: humans. This lack of understanding stymies
efforts to create safe, efficient, and productive human-machine teams. The Artificial
Social Intelligence for Successful Teams (ASIST) program seeks to develop founda-
tional AI theory and systems that demonstrate the basic machine social skills needed
to infer the goals and situational knowledge of human partners, predict what they will
need, and offer context-aware actions in order to perform as adaptable and resilient AI
teammates. ASIST performers will work to create agents that demonstrate a machine
ToM and the ability to participate in an effective team by representing and helping to

167
maintain shared models. The program seeks to create and employ a testbed for eval-
uating these agents using customizable open-world environments and standardized
interfaces. The interfaces will include a package of standard sensing channels (e.g.,
information streams from physical and virtual sensors that will be available to ASIST
agents) and communication/action channels (e.g., mechanisms for agents to convey
information to human teammates in the testbed and options they can use to engage
with the team). ASIST agents must operate in increasingly complex and specialized
environments; be adaptable to sudden perturbations in the mission or team, like the
loss of communication with a key teammate; and use noisy multi-channel observations
to represent the world and do complex inference and prediction.

168
Lasers for Universal Microscale Op-
tical Systems (LUMOS)
Dr. Gordon Keeler

https://www.darpa.mil/program/lasers-for-universal-microscale-optical-systems.html

GPT-3 Summary: The goal of the LUMOS program is to develop high-performance

lasers and amplifiers that can be integrated into manufacturable photonic platforms.
This will be achieved by heterogeneous integration of diverse materials. The three
vectors of focus are scaling complexity, scaling power, and scaling spectrum. The
end goal is to enableatomic microsystems for positioning, navigation, and timing
applications, as well as compact quantum sensors and information processing systems.

Full Details: Lasers have made a tremendous impact on our world – they are essen-
tial to diverse fields such as optical communications, remote sensing, manufacturing,
and medicine. At the same time, photonic integrated circuits have allowed unprece-
dented advances in optical systems for Department of Defense (DoD) applications,
including LiDAR, signal processing, chip-scale optical clocks, gyros, and data trans-
mission. However, these two technologies today are limited by the incompatibility of
the materials used to create them – silicon photonics are easy to manufacture but are
poor light emitters while compound semiconductors enable efficient emitters but are
difficult to scale for complex integrated circuits.The Lasers for Universal Microscale
Optical Systems (LUMOS) program aims to bring high performance lasers and ampli-
fiers to manufacturable photonics platforms through heterogeneous integration of di-
verse materials. LUMOS seeks to develop integrated photonics technology along three
vectors: scaling complexity, scaling power, and scaling spectrum. To achieve improved
scaling in complexity, LUMOS seeks to integrate thousands of lasers and amplifiers
with highly complex photonic integrated circuits for applications such as compact
optical phased array LiDAR and neuromorphic optical computing. To achieve power
scaling, LUMOS will aim to develop Watt-class lasers on a low-loss, high-speed pho-
tonics platform for radio frequency (RF) and microwave applications. For scaling
across the spectrum, LUMOS aims to create photonic circuits with integrated lasers
operating across the visible spectrum with a wavelength-by-design methodology to
enable atomic microsystems for positioning, navigation, and timing applications, as
well as compact quantum sensors and information processing systems.

169
Living Foundries
Dr. Anne Cheever

https://www.darpa.mil/program/living-foundries.html

GPT-3 Summary: GOAL: The Living Foundries program will enable the adapt-
able, scalable, and on-demand production of molecules by programming the funda-
mental metabolic processes of biological systems.
TECHNOLOGY METHODS: The program will include the development of next-
generation tools and technologies for engineering biological systems with the goal of
compressing the biological design-build-test-learn cycle.
IMPACT: This program will transform synthetic biomanufacturing into a predictable
engineering practice supportive of a broad range of national security objectives.

Full Details: Current and emerging Department of Defense (DoD) capabilities rely
upon access to a number of critical, high-value molecules that are often prohibitively
expensive, unable to be domestically sourced, and/or impossible to manufacture using
traditional synthetic approaches. DARPA’s Living Foundries program aims to enable
adaptable, scalable, and on-demand production of such molecules by programming
the fundamental metabolic processes of biological systems to generate a vast number
of complex molecules that are not otherwise accessible. Through Living Foundries,
DARPA is transforming synthetic biomanufacturing into a predictable engineering
practice supportive of a broad range of national security objectives. The program
includes two components: 1) Living Foundries: Advanced Tools and Capabilities for
Generalizable Platforms (ATCG), and 2) Living Foundries: 1000 Molecules. The first
component, Living Foundries: ATCG, completed successfully. It focused on the de-
velopment of next-generation tools and technologies for engineering biological systems
with the goal of compressing the biological design-build-test-learn cycle by at least
ten times in both time and cost, while increasing the complexity of systems that are
created. Technical areas of interest included design and automation tools, modular
genetic parts and devices, standardized test platforms and chassis, tools for rapid
physical construction of biological systems, editing and manipulation of genetic de-
signs, and new characterization and debugging tools for synthetic biological networks.
The ongoing Living Foundries: 1000 Molecules component seeks to further refine this
initial capability to significantly decrease the cost, improve the scalability, and ex-
pand the complexity of engineered systems for biomanufacturing. Efforts are focused
on using automation, novel genome editing tools, and machine learning technologies
to alleviate the challenges of prototyping. As a proof of concept, DARPA aims to
produce 1,000 molecules and material precursors spanning a wide range of defense-

170
relevant applications including industrial chemicals, pharmaceuticals, coatings, and
adhesives that can be customized to continuously evolving DoD needs while ensuring
continued leadership of the United States in the rapidly evolving field of synthetic
biology.

171
Safe Documents (SafeDocs)
Dr. Sergey Bratus

https://www.darpa.mil/program/safe-documents.html

GPT-3 Summary: The goal of the SafeDocs program is to develop verified

programming methodologies for building high assurance parsers for extant electronic
data formats, and novel methodologies for comprehending, simplifying, and reducing
these formats to their safe, unambiguous, verification-friendly subsets (”safe sub-
setting”).

The technology methods of this program include extracting the extant for-
mats’ de facto syntax (including any non-compliant syntax deliberately accepted
and substantially used in the wild); identifying a syntactically simpler subset of this
syntax that yields itself to use in verified programming while preserving the format’s
essential functionality; and creating software

Full Details: Today, code for input data validation is typically written manually in
an ad-hoc manner. For commonly-used electronic data formats, input validation is,
at a minimum, a problem of scale whereby specifications of these formats comprise
hundreds to thousands of pages. Input validation thus translates to thousands or
more conditions to be checked against the input data before the data can be safely
processed. Manually writing the code to parse and validate input, and then manu-
ally auditing whether that code implements all the necessary checks completely and
correctly, does not scale. Moreover, manual parser coding and auditing typically fails
even for electronic data formats specifically designed to be easier to perform such
tasks, e.g., JSON and XML. A variety of critical vulnerabilities have been found in
major parser implementations for these formats. Widely deployed mitigations against
crafted input attacks include (a) trying to prevent the flow of untrusted data to vul-
nerable software; and (b) testing software with randomized inputs to find and patch
flaws that could be triggered by maliciously created inputs. Unfortunately, neither of
these approaches offer security assurance guarantees.Mitigations for preventing the
flow of untrusted data to vulnerable software, which can be implemented via network
or host-based measures such as firewalls, application proxies, antivirus scanners, etc.,
neither remove the underlying vulnerability from the target, nor encode complete
knowledge of document or message format internals. Attacker bypasses of such mit-
igations exploit incompleteness of the mitigations’ understanding of the data format
to exploit the still-vulnerable targets.The Safe Documents (SafeDocs) program will
develop novel verified programming methodologies for building high assurance parsers

172
for extant electronic data formats, and novel methodologies for comprehending, sim-
plifying, and reducing these formats to their safe, unambiguous, verification-friendly
subsets (“safe sub-setting”). SafeDocs will address the ambiguity and complexity ob-
stacles that hinder the application of verified programming posed by extant electronic
data formats. SafeDocs’ multi-pronged approach will combine:extracting the extant
formats’ de facto syntax (including any non-compliant syntax deliberately accepted
and substantially used in the wild);identifying a syntactically simpler subset of this
syntax that yields itself to use in verified programming while preserving the format’s
essential functionality; andcreating software construction kits for building secure, ver-
ified parsers for this syntactically simpler subset, and high-assurance translators for
converting extant instances of the format to this subset. The parser construction
kits developed by SafeDocs will be usable by industry programmers who understand
the syntax of electronic data formats but lack the theoretical background in verified
programming. These tools will enable developers to construct verifiable parsers for
new electronic data formats as well as extant ones. The tools will guide the syntactic
design of new formats, by making verification-friendly format syntax easy to express,
and vice versa.Additional information is available in the SafeDocs BAA.

173
Nascent Light-Matter Interactions (NLM)
Dr. Rohith Chandrasekar

https://www.darpa.mil/program/nascent-light-matter-interactions.html

GPT-3 Summary: The goal of the NLM program is to develop new models that can
describe and predict novel light-matter interactions. These models will provide design
tools and performance limits for new engineered light-matter interactions. Important
applications addressed in the program include new sources, non-reciprocal behavior,
parametric phenomena, limiters, electromagnetic drives, and energy harvesting.

Full Details: Recent advances in our understanding of light-matter interactions, of-

ten with patterned and resonant structures, reveal nascent concepts for new interac-
tions that may impact many applications. Examples of these novel phenomena include
interactions involving active media, symmetry, non-reciprocity, and linear/nonlinear
resonant coupling effects. Insights regarding the origins of these interactions have the
potential to transform our understanding of how to control electromagnetic waves
and design for new light-matter interactions. The goal of NLM is to bring together
and integrate these emerging phenomena with fundamental models that can describe
and predict new functionality. These models will provide design tools and delineate
the performance limits of new engineered light-matter interactions. Important appli-
cations to be addressed in the program include synthesizing new material structures
for sources, non-reciprocal behavior, parametric phenomena, limiters, electromagnetic
drives, and energy harvesting.

174
Operational Fires (OpFires)
Lt. Col. Joshua Stults

https://www.darpa.mil/program/operational-fires.html

GPT-3 Summary: GOAL: The Operational Fires (OpFires) program will develop
a ground-launched system to rapidly and precisely engage critical time sensitive
targets with hypersonic boost glide weapons.

TECHNOLOGY METHODS: Conduct a series of subsystem tests to evaluate

component design and system compatibility. Test the system with integrated
end-to-end flight tests.

IMPACT: Enable ground forces to rapidly and precisely engage critical time
sensitive targets with hypersonic boost glide weapons.

Full Details: The goal of the Operational Fires (OpFires) program is to develop
and demonstrate a novel ground-launched system enabling hypersonic boost glide
weapons to penetrate modern enemy air defenses and rapidly and precisely engage
critical time sensitive targets.OpFires seeks to develop an advanced booster capable
of delivering a variety of payloads at a variety of ranges. Additional considerations
include the need for compatible mobile ground launch platforms enabling integration
with existing ground forces and infrastructure, and specific system attributes required
for rapid deployment and redeployment. The OpFires program will conduct a series
of subsystem tests designed to evaluate component design and system compatibility,
and culminate in integrated end-to-end flight tests.

175
PReemptive Expression of Protec-
tive Alleles and Response Elements
(PREPARE)
Dr. Amy Jenkins

https://www.darpa.mil/program/preemptive-expression-of-protective-alleles-and-response-
elements.html

GPT-3 Summary: The goal of the PREPARE program is to develop new

medical countermeasures (MCMs) that can quickly and effectively protect against a
broad range of threats, including natural disasters, accidents, pandemics, and attacks.

The program will seek to identify and validate protective genes, develop novel
MCMs that can specifically and multiplexably control these genes, and create
methods for in vivo delivery of these MCMs to confer protection in healthy animal
models.

If successful, this program could have a major impact on public health and
national security by providing a way to rapidly and effectively respond to a wide
range

Full Details: Pathogens with pandemic potential, toxic chemicals, and radioactive
materials all endanger public health and pose a threat to national security. Despite
investment in the development of medical countermeasures (MCMs) to address these
threats, many existing MCMs suffer from limited applicability, insufficient efficacy,
requirements for repeat dosing, lengthy and complex manufacturing processes, and
logistically burdensome storage requirements. In many cases, unique threats require
unique responses, setting up a “one threat, one MCM” paradigm. These limita-
tions make it extraordinarily difficult to treat the large numbers of individuals who
could be impacted by the outcomes of a natural disaster, accident, pandemic out-
break, or directed attack in a heavily populated area. The PReemptive Expression
of Protective Alleles and Response Elements (PREPARE) program aims to develop
a new approach by identifying innate host genetic defenses against threats and de-
veloping novel MCMs that can quickly activate and modulate these genes to boost
protection—without altering the underlying genetic code. The programmable, gene-
modulating MCMs DARPA envisions are intended to provide timely and robust pro-
tection against a broad range of threats; however, within the PREPARE program,

176
research teams will seek to demonstrate proof of concept by addressing three ma-
jor threat areas: influenza infection, organophosphate poisoning, and exposure to
gamma irradiation.PREPARE researchers will apply new insights from the fields of
genome editing and in vivo delivery to develop safe, programmable, and transient
treatments that precisely tune the activity of protective genes to preemptively acti-
vate and strengthen a host’s intrinsic defenses. Work on the program will identify and
validate protective genes; develop novel MCMs capable of specific and multiplexable
control of endogenous protective genes; and create methods to enable tissue-specific,
in vivo delivery of programmable MCMs to confer protection in healthy animal mod-
els for relevant durations. The integration of these aims will support a capability to
generate programmable, gene-modulating MCMs with a path towards clinical trans-
lation.

177
Radio Frequency Machine Learning
Systems (RFMLS)
Mr. John Davies

https://www.darpa.mil/program/radio-frequency-machine-learning-systems.html

GPT-3 Summary: GOAL: The goal of the program is to develop the foundations
for applying modern data-driven Machine Learning (ML) to the RF Spectrum
domain.

TECHNOLOGY METHODS: The program will focus on developing software

that can learn RF features, synthesize waveforms, understand the difference between
important and unimportant signals, and configure hardware resources.

IMPACT: The program could improve the performance of traditional RF sys-

tems such as radar, signals intelligence, electronic warfare, and communications.

Full Details: The goal of the Radio Frequency Machine Learning Systems (RFMLS)
Program is to develop the foundations for applying modern data-driven Machine
Learning (ML) to the RF Spectrum domain. These innovations form the basis of a
new wave of Signal Processing technologies to address performance limitations of con-
ventionally designed radio frequency (RF) systems such as radar, signals intelligence,
electronic warfare, and communications. Over the last decade ML has been ap-
plied successfully to numerous sensor modalities, and is now common place in many
commercial applications, including object and facial recognition in images, speech
recognition in acoustic signals, and text parsing and reasoning from documents. Key
to each of these innovations was the evolution from hand-engineered approaches tai-
lored to each problem, to solutions that learned from large datasets. RF systems
conversely are still designed using models and equations based on idealized assump-
tions and approximations regarding hardware, environment, and the problem being
solved. The inaccuracy of these assumptions challenge our ability to perform tasks
such as identification of signals among the ever-increasing myriad which populate the
wireless landscape. Under the program, RFMLS systems will seek to learn to perform
four specific tasks. Each task emphasizes a core constituent capability of RF ML. The
four solutions can be combined and applied to address DoD operational needs in the
RF Spectrum. RF Fingerprinting. Traditional wireless security relies on a software
“identity” for each wireless device, which can often be hacked or otherwise cloned.
The RFMLS system will aim to learn to recognize a specific transmitter based on

178
the unique RF fingerprint naturally imparted by hardware imperfections within that
transmitter. This task focuses on learning RF features. RF Fingerprint Enhancement.
To further enhance wireless security, a communication system learns to modify it’s
transmit waveforms to enhance its natural fingerprint. This task focuses on learning
to synthesize waveforms. Spectrum Awareness. Traditional systems which monitor
the RF spectrum use narrow bandwidths and relatively simple strategies (such as
the frequency of transmission) to identify the signals occupying the wireless spec-
trum. Availability of commodity analog-to-digital converters with wide bandwidths
combined with proliferation of software defined radios, spectrum sharing, and general
wireless technology growth, challenge these approaches. RFMLS systems will learn
to understand the difference between important and unimportant signals present in
large bandwidths in order to build more useful and accurate spectrum awareness.
This task emphasizes goal-driven attention. Autonomous RF System Configuration.
To further enhance spectrum awareness performance, a RFMLS system will seek to
learn how best to tune and configure its hardware resources in order to maximize the
number of important signals discovered in harsh RF environments. This task stresses
hardware configuration and control.

179
Extreme Optics and Imaging (EX-
TREME)
Dr. Rohith Chandrasekar

https://www.darpa.mil/program/extreme-optics-and-imaging.html

GPT-3 Summary: The EXTREME program’s goal is to develop new optical com-
ponents, devices, systems, architectures and design tools using Engineered Optical
Materials (EnMats) to enable new functionality and/or vastly improve size, weight,
and power characteristics of traditional optical systems.

Full Details: The goal of the EXTREME Program is to develop new optical com-
ponents, devices, systems, architectures and design tools using Engineered Optical
Materials (EnMats) to enable new functionality and/or vastly improve size, weight,
and power characteristics of traditional optical systems. EnMats are broadly defined
to include, but are not limited to, metamaterials (both metallic and dielectric), scat-
tering surfaces and volumes, holographic structures, and diffractive elements. Early
examples of EnMats have been used to design and build multifunctional elements and
to dynamically control light, seemingly going beyond standard “laws” of reflection and
refraction. The EXTREME program will explore this optical design space and aims
to understand the trade-offs, and harness the possibilities, afforded by EnMats.

180
Systematizing Confidence in Open
Research and Evidence (SCORE)
Dr. Bartlett Russell

https://www.darpa.mil/program/systematizing-confidence-in-open-research-and-evidence.html

GPT-3 Summary: GOAL: The SCORE program will develop and deploy au-
tomated tools to assign quantitative ”confidence scores” to different social and
behavioral science (SBS) research results and claims.

METHODS: These tools will assign explainable confidence scores with a relia-
bility that is equal to, or better than, the best current human expert methods.

IMPACT: If successful, SCORE will enable DoD personnel to quickly cali-

brate the level of confidence they should have in the reproducibility and replicability
of a given SBS result or claim, and thereby increase the effective use of SBS literature
and

Full Details: The Department of Defense (DoD) often leverages social and be-
havioral science (SBS) research to design plans, guide investments, assess outcomes,
and build models of human social systems and behaviors as they relate to national
security challenges in the human domain. However, a number of recent empirical
studies and meta-analyses have revealed that many SBS results vary dramatically
in terms of their ability to be independently reproduced or replicated, which could
have real-world implications for DoD’s plans, decisions, and models. To help address
this situation, DARPA’s Systematizing Confidence in Open Research and Evidence
(SCORE) program aims to develop and deploy automated tools to assign ”confidence
scores” to different SBS research results and claims. Confidence scores are quantita-
tive measures that should enable a DoD consumer of SBS research to understand the
degree to which a particular claim or result is likely to be reproducible or replicable.
These tools will assign explainable confidence scores with a reliability that is equal to,
or better than, the best current human expert methods. If successful, SCORE will
enable DoD personnel to quickly calibrate the level of confidence they should have
in the reproducibility and replicability of a given SBS result or claim, and thereby
increase the effective use of SBS literature and research to address important human
domain challenges, such as enhancing deterrence, enabling stability, and reducing
extremism.

181
CONverged Collaborative Elements
for RF Task Operations (CONCERTO)
Dr. Kevin Rudd

https://www.darpa.mil/program/converged-collaborative-elements-for-rf-task-operations.html

GPT-3 Summary: The goal of the CONCERTO program is to develop an intelli-

gent resource management system to coordinate RF mission goals in spectrum, space,
and time. The resource manager will use the latest Reinforcement Learning and Ge-
netic Algorithms to efficiently distill the vast option space in real time. The modular
architecture is also scalable across a range of platforms and will be demonstrated on
third party payloads in a flight test at the conclusion of the program.

Full Details: Dominance of the radio frequency (RF) spectrum is critical to suc-
cessful U.S. military operations. Today, we do this using discrete radar, electronic
warfare (EW), and communication payloads that are separately designed, procured,
and integrated on platforms. These payloads typically use dedicated apertures, are
realized with tightly coupled hardware and software, and are not well-coordinated
in their use of spectrum. This rigid and constrained approach makes it difficult
and time-consuming to adopt new technology, adapt to rapidly changing adversary
threats, maneuver functions effectively in spectrum, and create comprehensive com-
pact RF systems. Finally, such converged, flexible, multi-function systems are accom-
panied by an overwhelming growth in the number of options due to the anomaly of
cheap complexity. The CONverged Collaborative Elements for RF Task Operations
(CONCERTO) program represents the ultimate software defined RF tile for Mosaic
Warfare. The program addresses these challenges by developing a modular architec-
ture for adaptive, converged RF systems to realize intelligent control of converged
RF payloads supporting radar, EW, and communications simultaneously. The CON-
CERTO systems under development realize multi-function operations in less space
and power than the combined collections of discrete systems, maximize the use and
overlapping capabilities of common apertures, and speed technology migration by
decoupling hardware from software and firmware. More specifically, CONCERTO is
developing a multi-objective resource management system to intelligently coordinate
RF mission goals in spectrum, space, and time to maximize the effectiveness of multi-
ple complementary converged payloads. Leveraging the latest Reinforcement Learn-
ing and Genetic Algorithms, the CONCERTO resource manager is able to efficiently
distill the vast option space in real time dramatically outperforming the traditional
greedy approaches. Finally, the modular architecture is also scalable across a range

182
of platforms and will be demonstrated on third party payloads in a flight test at the
conclusion of the program.

183
Microscale Plasma Devices (MPD)
(Archived)

https://www.darpa.mil/program/microscale-plasma-devices.html

GPT-3 Summary: GOAL: The DARPA Microscale Plasma Devices (MPD)

program seeks to enable a revolutionary new class of electronics and signal processing
devices based on microscale plasma cavities capable of withstanding damaging elec-
tromagnetic pulses (EMP) and of operating in extreme temperature and radiation
environments.

TECHNOLOGY METHODS: This program includes two research objectives:

1) To advance fundamental understanding of microscale plasma science and device
operations for one or more key MPD design parameters, such as gas type or pressure;
and 2) To design, model and fabricate microscale plasmas devices that target
advances in specified

Full Details: Many defense electronics are susceptible to radiation and high tem-
peratures. Developing electronics that can withstand harsh conditions would expand
the types of environments in which DoD electronics may be used. The DARPA Mi-
croscale Plasma Devices (MPD) program seeks to enable a revolutionary new class
of electronics and signal processing devices based on microscale plasma cavities ca-
pable of withstanding damaging electromagnetic pulses (EMP) and of operating in
extreme temperature and radiation environments. MPD will work to develop small
feature size (< 20 micrometer) devices for use in electronic protection systems for
Defense applications. Plasma devices sought should be capable of switching high-
carrier densities (1018/cm3 ) with speeds of 100ps or faster. MPD seeks to develop
dual- and multi-terminal devices and demonstrate them in circuits, substrates and
materials—highlighting the efficacy of the proposed approaches as well as the unique
benefits of microplasma-based systems. This program includes two research objec-
tives: To advance fundamental understanding of microscale plasma science and device
operations for one or more key MPD design parameters, such as gas type or pressure.
To design, model and fabricate microscale plasmas devices that target advances in
specified performance objectives. MPD also aims to develop modeling, simulation
and design tools for rapid manufacturing of devices. Work under MPD began in July
2011.

184
Automating Scientific Knowledge Ex-
traction (ASKE)
Dr. Joshua Elliott

https://www.darpa.mil/program/automating-scientific-knowledge-extraction.html

GPT-3 Summary: GOAL: ASKE will develop technology to automate some of the
manual processes of scientific knowledge discovery, curation and application.
TECHNOLOGY METHODS: Develop approaches to make it easier for scientists to
build, maintain and reason over rich models of complex systems.
IMPACT: This could make it easier for scientists to understand and work with com-
plex systems.

Full Details: The Automating Scientific Knowledge Extraction (ASKE) program

aims to develop technology to automate some of the manual processes of scientific
knowledge discovery, curation and application. ASKE is part of DARPA’s Artificial
Intelligence Exploration (AIE) program, a key component of the agency’s broader
AI investment strategy aimed at ensuring the United States maintains an advantage
in this critical and rapidly accelerating technology area. ASKE seeks to develop ap-
proaches to make it easier for scientists to build, maintain and reason over rich models
of complex systems – which could include physical, biological, social, engineered or
hybrid systems – by interpreting and exposing scientific knowledge and assumptions
in existing model code and documentation, identifying new data and information
resources automatically, extracting useful information from these sources, integrat-
ing this useful information into machine-curated expert models, and executing these
models in robust ways.

185
Blackjack
Mr. Stephen Forbes

https://www.darpa.mil/program/blackjack.html

GPT-3 Summary: The Blackjack program seeks to develop a global high-speed

network in low Earth orbit (LEO) to provide the Department of Defense with highly
connected, resilient, and persistent coverage. The program objectives are to develop
payload and mission-level autonomy software and to demonstrate payloads in LEO to
augment NSS assets. The program will use advanced commercial manufacturing for
military payloads and the spacecraft bus. The goal is to show LEO performance that
is on par with current systems in geosynchronous orbit with the spacecraft combined
bus, payload(s), and launch costs under $6 million per orbital node while the payload

Full Details: National Security Space (NSS) assets, critical to U.S. warfighting ca-
pabilities, traditionally reside in geosynchronous orbit to deliver persistent overhead
access to any point on the globe. In the increasingly contested space environment,
these exquisite, costly, and monolithic systems have become vulnerable targets that
would take years to replace if degraded or destroyed. DARPA’s Blackjack program
aims to develop and demonstrate the critical elements for a global high-speed network
in low Earth orbit (LEO) that provides the Department of Defense with highly con-
nected, resilient, and persistent coverage. Blackjack seeks to incorporate commercial
sector advances in LEO, including design of LEO constellations intended for broad-
band internet service, of which the design and manufacturing could offer economies of
scale previously unavailable. DARPA is interested in capitalizing on these advances
to demonstrate military utility, emphasizing a commoditized bus and low-cost inter-
changeable payloads with short design cycles and frequent technology upgrades. The
key program objectives are: Develop payload and mission-level autonomy software
and demonstrate autonomous orbital operations including on-orbit distributed de-
cision processors. Develop and implement advanced commercial manufacturing for
military payloads and the spacecraft bus. Demonstrate payloads in LEO to augment
NSS assets. The driver will be to show LEO performance that is on par with current
systems in geosynchronous orbit with the spacecraft combined bus, payload(s), and
launch costs under $6 million per orbital node while the payloads meet size, weight,
and power constraints of the commercial bus.

186
COmpact Front-end Filters at the
ElEment-level (COFFEE)
Dr. Benjamin Griffin

https://www.darpa.mil/program/compact-front-end-filters-at-the-element-level.html

GPT-3 Summary: The goal of the COmpact Front-end Filters at the ElEment-level
(COFFEE) program is to develop a new class of integrable, high-frequency RF filters
for next-generation wideband arrays. The COFFEE filter technology will address the
combination of size, performance, and reproducibility to enable protection at every
element of a wideband AESA. COFFEE aims to build integrable filters that operate
over a wide range of frequencies that are also small enough to fit behind each element
of the phased array thus making the array more robust and resistant to interference

Full Details: Interest in wideband Active Electronically Scanned Array AESAs

with digital-at-every-element architectures has been increasing over the past decade.
Wideband AESAs are more versatile and robust because they allow for operation at
different frequencies, as opposed to narrowband arrays that are tuned for a specific
use and frequency. Digital-at-every-element architectures allow AESAs to perform
beamforming in the digital domain, enabling the simultaneous collection of many
beams. However, current implementations of wide bandwidth receivers often have
limited dynamic range that leave them vulnerable to electronic interference and jam-
ming, particularly in congested RF environments. Further, digital-at-every-element
requires filtering at the element level, leaving very little room to integrate conventional
filter technologies.To address the challenges impeding the use of wideband AESAs in
congested RF environments, DARPA is executing the COmpact Front-end Filters
at the ElEment-level (COFFEE) program. COFFEE aims to develop a new class
of integrable, high-frequency RF filters for next-generation wideband arrays. The
COFFEE filter technology will address the combination of size, performance, and
reproducibility to enable protection at every element of a wideband AESA. COFFEE
aims to build integrable filters that operate over a wide range of frequencies that are
also small enough to fit behind each element of the phased array thus making the
array more robust and resistant to interference. Key to the program is the develop-
ment of filter technology that can address all microwave frequencies of a wideband
AESA’s bandwidth without sacrificing performance. Further, the target filters must
be physically small compared to the element area as the available space for element-
level integration decreases significantly as AESA bandwidth increases. Finally, to
ensure uniformity the COFFEE filters should be manufacturable with reproducible

187
performance at each of the array elements.A main focus of the research will be on
developing a new class of resonators and integrable microwave filters that address
COFFEE’s technical objectives. In addition, research studies into compact mm-wave
resonators will be conducted to inform new technical approaches for potential fu-
ture efforts for integrable mm-wave filters.COFFEE is a part of DARPA’s Electronics
Resurgence Initiative (ERI) – a five-year, $1.5 billion investment in the advancement
of the U.S. semiconductor industry. The program addresses a key ERI focus area of
revolutionizing communications for the 5G era and beyond.

188
Assessing Immune Memory (AIM)
Dr. Kristen Jordan

https://www.darpa.mil/program/assessing-immune-memory.html

GPT-3 Summary: GOAL: The AIM program will study how to create long-lasting
immunity in humans through vaccination.
METHODS: Research will take a systems-level view of the response to vaccination
and dissect the mechanisms that lead to long lasting protection.
IMPACT: The program will create a tool to predict vaccine duration of protection and
the associated mechanisms, without waiting years for retrospective determination.

Full Details: Military service members rely on effective vaccination for the preven-
tion of communicable disease as well as to guard against biothreat exposure. Many
current vaccines lack durability (i.e., do not provide effective protection over long
periods of time), and there are multiple pathogens and threats that lack prophylactic
options altogether. The current state of vaccine durability assessment is to take a
“wait and see” approach, largely owing to ignorance of mechanisms underlying im-
mune memory, as well as an inability to measure the cellular contributors that invoke
long-lasting immune protection. Formation of immune memory is a complex physio-
logical process characterized by a diverse array of cellular interactions and signaling
processes. The Assessing Immune Memory (AIM) program seeks to determine early
on if a vaccine candidate will later provide long-lasting immune protection in humans,
a current impossibility that would benefit the warfighter and nation immensely. To
accomplish this goal, AIM will take a systems-level view of the response to vaccination
and dissect the mechanisms that lead to long lasting protection. This systems-level
understanding will then be implemented as a tool that can predict vaccine duration
of protection and the associated mechanisms without waiting years for retrospective
determination.

189
Agile Teams (A-Teams)
Dr. Bartlett Russell

https://www.darpa.mil/program/agile-teams.html

GPT-3 Summary: GOAL: The A-Teams program seeks to create a systematic

methodology to design teams that best use the capabilities of both humans and
machines and that can achieve enhanced performance in uncertain, dynamic, and
co-evolving environments.
METHODS: The A-Teams program will use experimental testbeds to validate new
abstractions for human-machine team architectures.
IMPACT: The A-Teams program could lead to the development of more effective and
efficient human-machine teams that are better equipped to handle complex tasks in
uncertain and dynamic environments.

Full Details: The Agile Teams (A-Teams) program aims to discover, test, and
demonstrate generalizable mathematical abstractions for the design of agile human-
machine teams and to provide predictive insight into team performance. While
human-machine teams have been the subject of considerable past work in artificial
intelligence and autonomy, designing agile team architectures remains largely a trial-
and-error enterprise. The A-Teams program seeks to create a systematic methodology
to design teams that best use the capabilities of both humans and machines and that
can achieve enhanced performance in uncertain, dynamic, and co-evolving environ-
ments. These new abstractions will be validated using experimental testbeds aimed
to support reproducible evaluation of human-machine team architectures in a diverse
range of problem contexts.

190
Posh Open Source Hardware (POSH)
Dr. James Wilson

https://www.darpa.mil/program/posh-open-source-hardware.html

GPT-3 Summary: The POSH program’s goal is to create a more efficient and secure
way to develop complex system-on-chip (SoC) designs. The program will achieve
this by developing new methodologies for verification and mathematical inspection
of analog and digital IP blocks. These blocks will be released on an open source
platform, serving as a foundation for rapid design of complex secure SoCs. The
POSH program’s impact is to reduce the cost and time associated with developing
new SoCs.

Full Details: The unrelenting progression of Moore’s Law has created a steady
cadence to ever-smaller transistors and more powerful chips, allowing billions of tran-
sistors to be integrated on a single system-on-chip (SoC). However, engineering pro-
ductivity has not kept pace with Moore’s Law, leading to prohibitive increases in
development costs and team sizes for leading-edge SoC design. To help manage the
complexity of SoC development, design reuse in the form of Intellectual Property (IP)
modules has become the primary strategy. IP modules are pre-designed, functional
circuit blocks that are similar in nature to software library functions and are developed
internally by an organization or procured from an external, third-party IP vendor.
Defense systems today may leverage over a hundred unique IP blocks, which creates
significant productivity improvements but also limits visibility into the final SoC’s
behavior and offers no clear path for complete security or integrity verification due to
schedule and cost pressures. Further, the single layer point-to-point approach asso-
ciated with the current IP licensing model limits the scope of reuse and abstraction.
A candidate approach for creating a deeply layered, transparent SoC development
model is to draw from best practices in the software community and leverage open
source hardware IP. However, the idea of reusing the open source IP available today
has not seen wide adoption due to the high cost of failure, limited capabilities, and
inability to prove functionality prior to adoption. To help address the current design
complexity challenges, the Posh Open Source Hardware (POSH) program seeks to
enable mathematically provable secure electronics and create an open source hard-
ware IP ecosystem, along with accompanying validation tools. Under the program,
researchers will work to develop methodologies, standards, and simulation as well
as emulation technologies for the verification and mathematical inspection of analog
and digital IP to provide proof of functionality and security. The program also aims
to develop and release a number of silicon-proven analog and digital IP blocks on an

191
open source platform to serve as a foundation for rapid design of complex secure SoCs
at leading edge technology nodes.

192
Gamebreaker
Dr. Greg Kuperman

https://www.darpa.mil/program/gamebreaker.html

GPT-3 Summary: The Gamebreaker program seeks to use Artificial Intelligence

(AI) to improve game balance in existing video games. The goal is to develop a
methodology to empirically derive an accurate model for game balance, which can
then be used to assess and recommend modifications to the game that will result
in an imbalanced condition. This could help the military develop winning strategies
when the adversary’s objectives are not clearly known.

Full Details: The Gamebreaker program seeks to develop and apply Artificial Intel-
ligence (AI) to existing open-world video games to quantitatively assess game balance,
identify parameters that significantly contribute to balance, and explore new capa-
bilities, tactics, and rule modifications that are most destabilizing to the game. For
Gamebreaker, game balance is defined as an inherent property of the game that re-
flects the win/loss ratio of players of equal skill level based on strategies and tactics
employed within the game. For example, in a balanced game, if the skill level of
both players is equal, each player will win approximately 50 percent of the total
games played. Similarly, if the skill level of both players is equal but one player
wins disproportionately due to inherent advantages arising from a condition of the
game construct, the game is unbalanced. The commercial gaming industry has a
long-standing interest in maintaining game balance since balanced games are more
entertaining, and market pressures drive their development. Instead, the military has
the opposite goal whereby strategists deliberately explore technologies that maximize
imbalance. While AI techniques have demonstrated the ability to master games of
increasing complexity, there has been very little research in the application of AI
to game modification itself. Complex, multi-domain modeling and simulation envi-
ronments currently under development by several DARPA programs aim to create
useful “Mosaic” models within which to experiment on new warfighting constructs
using distributed, adaptive, all-domain force composition, tactics, and strategies –
yet these models do not currently exist. It is reasonable to assume, however, that
once these simulation environments reach maturity, an “AlphaMosaic” equivalent will
be capable of searching for optimal strategies and tactics in the same way AlphaGo
and AlphaStar agents have already proven effective in exploiting their respective
game environments. None of the existing efforts plan take into account modifica-
tions to the construct of the game itself and represents the fundamental motivation
for the Gamebreaker program.Gamebreaker performers will begin by first develop-

193
ing a methodology to empirically derive an accurate model for game balance using
a variety of algorithmic approaches applied to several different games. Ideally these
models will result in successful approaches to arrive at an effective game balance state
equation which can then be used to assess and recommend modifications to the game
that will result in an imbalanced condition. In future conflicts, DoD investment is
designed to maximize imbalance to create an advantage or to seek equilibrium when
an adversary is seeking an advantage. New AI Algorithms inspired by Gamebreaker
could help develop winning warfighting strategies when the adversary’s objectives –
i.e. the “rules of the game” – are not clearly known. By exploiting game balance,
Gamebreaker addresses an existing gap in AI and data analytics research as applied
to current wargaming and simulation.

194
Harnessing Autonomy for Counter-
ing Cyberadversary Systems (HACCS)
Ms. Allison Kline

https://www.darpa.mil/program/harnessing-autonomy-for-countering-cyberadversary-
systems.html

GPT-3 Summary: GOAL: Develop a system to effectively counter botnets and

other large-scale malware threats.
TECHNOLOGY METHODS: Develop software to automatically identify and neu-
tralize botnets.
IMPACT: Improve security posture by effectively countering botnets and other mal-
ware threats.

Full Details: Malicious actors are currently able to compromise and use with im-
punity large numbers of devices owned and operated by third parties. Such collections
of compromised and conscripted devices, commonly referred to as botnets, are used
for criminal, espionage, and computer network attack purposes (often a combination
of all three). Recent examples of botnets and similar malicious code include Mirai,
Hidden Cobra, WannaCry, and Petya/NotPetya. The potential scale of their effects
make such malware a national security threat. The May 11, 2017, Presidential Ex-
ecutive Order on Strengthening the Cybersecurity of Federal Networks and Critical
Infrastructure specifically identifies botnets as a high priority national security issue.
Improving the security posture of Department of Defense (DoD) networks alone is
insufficient to counter such threats to national security, as the majority of botnet
nodes reside in neutral networks (“gray space”). Current incident response meth-
ods are too resource- and time-consuming to address the problem at scale. Active
defense methods are insufficiently precise and predictable in their behavior, posing
a risk that they may cause processing issues or other side effects. What is needed
is the ability to identify and neutralize botnets and other large-scale malware from
compromised devices and networks in a scalable, timely, safe, and reliable manner,
in accordance with appropriate privacy and other legal authorities. To achieve the
necessary scale and timeliness, such a capability must be effective even if the own-
ers of botnet conscripted networks are unaware of the infection and are not actively
participating in the neutralization process.The HACCS program will investigate the
feasibility of creating safe and reliable autonomous software agents that can effectively
counter malicious botnet implants and similar large-scale malware. The program will
do so by developing a quantitative framework and established parameters for their

195
safe, reliable, and effective use. HACCS performers will develop the techniques and
algorithms necessary to measure the accuracy of identifying botnet-infected networks,
the accuracy of identifying the type of devices residing in a network, and the stability
of potential access vectors. The program will take an experimental approach to verify
the implementation of such autonomous agents and the rules under which they oper-
ate, and to measure the effectiveness of denying, degrading, and disrupting botnets
and individual botnet implants without affecting the systems and networks on which
they reside.The HACCS program seeks to:Accurately identify and fingerprint botnet-
conscripted networks to determine the presence of botnet implants, the number and
types of devices present on said networks, and the software services running on these
devices with sufficient precision to infer the presence of known vulnerabilities (also
referred to as “n-day” vulnerabilities); Generate non-disruptive software exploits for
a large number of known vulnerabilities that can be used to establish initial presence
in each botnet-conscripted network without affecting legitimate system functional-
ity; andCreate high-assurance software agents that safely, reliably, and autonomously
navigate within botnet-conscripted networks, identify botnet implants, and neutral-
ize them or otherwise curtail their ability to operate, while minimizing side effects to
these neutral systems and infrastructure.

196
Recovery of Symbolic Mathematics
from Code (ReMath)
Dr. Sergey Bratus

https://www.darpa.mil/program/recovery-of-symbolic-mathematics-from-code.html

GPT-3 Summary: GOAL:

The ReMath program seeks to discover whether recent advances in artificial

intelligence (AI) techniques can effectively recover mathematical structures imple-
mented in software into their natural mathematical forms of symbolic expression.

METHODS:

The program will use techniques such as neural machine translation, sequence-
to-sequence encoders, and others to try and recover mathematical structures from
software code.

IMPACT:

If successful, the program could improve the understanding of complex software and
enable future methods for analysis and testing of cyber physical systems.

Full Details: Extracting symbolic representations of software’s algorithmic parts,

such as control laws for a physical process encoded in a cyber-physical system, cur-
rently requires fully manual analysis by highly specialized experts. There is no mech-
anized capability to translate and route relevant parts of the software to experts,
such as control engineers, in an appropriate form for them to effectively analyze the
mathematical expressions. In contrast, malware analysis has become considerably
automated with aspects, such as provenance and behavioral characterization, gaining
considerable traction in recent years.Recent research demonstrates effective use of
artificial intelligence (AI) translation techniques for tasks like mechanized symbolic
differentiation and integration, in which going in one direction is relatively straightfor-
ward. Inverse problems like symbolic integration, however, are more challenging and
require expert ingenuity in repeated recognition and application of patterns.Although
programming mathematical models is not as straightforward as symbolic differenti-
ation, it is still a relatively straightforward skill with massive quantities of available
examples and broadly taught practices. Conversely, recovering a mathematical de-

197
scription of software, even when the function of the software is known, has defied
automation.The Recovery of Symbolic Mathematics from Code (ReMath) Artificial
Intelligence Exploration (AIE) program aims to discover whether a combination of
recent advances in AI techniques, such as neural machine translation, sequence-to-
sequence encoders, etc., can effectively recover mathematical structures implemented
in software into their natural mathematical forms of symbolic expression. These tech-
niques could improve the understanding of complex software and may enable future
methods for analysis and testing of cyber physical systems.For more information, see
the ReMath Program Announcement.

198
Knowledge Management at Scale and
Speed (KMASS)
Mr. Ted Senator

https://www.darpa.mil/program/knowledge-management-at-scale-and-speed.html

GPT-3 Summary: GOAL: The KMASS program will enable the effective use and
documentation of knowledge so it can be applied in a variety of contexts.

TECHNOLOGY METHODS:

The KMASS program will research, develop, integrate, evaluate, and demon-
strate underlying technology that will enable effective use of documented knowledge,
acquisition of new knowledge as part of regular workflows, and application of useful
knowledge when and where it is required and with necessary granularity.

IMPACT:

The goal of the KMASS program is to deliver user specific knowledge “nuggets” that
are useful for

Full Details: Organizations, including the military, store massive amounts of knowl-
edge intended for human consumption, e.g., doctrine, policy, procedures, etc. Creat-
ing these documents, videos, and other modes is both expensive and time consuming.
They are not structured or indexed to support rapid and appropriate application
to particular tasks and may be inconsistent or confusing. Further, this knowledge
must be augmented by local and timely knowledge appropriate to the particulars of
a task, accounting for any unique context or situations. Capturing knowledge along
with this context as it is created currently requires a dedicated effort on the part
of the person performing the task, which often does not happen. Applying stored
knowledge requires a user to know that it exists, where it exists, when it is needed
and relevant, how to retrieve it, and how to locate the specifics in what typically
is a multi-page document or several minute long video with audio. Applying stored
knowledge may not always be possible given time constraints.The Knowledge Man-
agement at Scale and Speed (KMASS) program aims to research, develop, integrate,
evaluate, and demonstrate underlying technology that will enable effective use of doc-
umented knowledge, acquisition of new knowledge as part of regular workflows, and
application of useful knowledge when and where it is required and with necessary

199
granularity. KMASS technology seeks to scale to a broad set of tasks and contexts
across an organization by collecting and modifying knowledge “in-the-flow” as part
of regular task execution and applying the knowledge documented for one purpose
to other purposes as appropriate. The goal of the KMASS program is to deliver user
specific knowledge “nuggets” that are useful for a current task—whether the knowl-
edge is requested or not by the user—exactly when needed, while avoiding irrelevant
or already known information. This concept is a core tenet of KMASS and may
be referred to as the “JustINs” – i.e., just in time, just enough, and just for me.1
KMASS systems also will contain a persistent knowledge store comprising source
documents in human understandable form in multiple modalities (e.g., text, videos,
presentations, etc.), augmented with appropriate tags and indexed for identification,
retrieval, linking, and application that will update at the speed of task performance.
KMASS requires advances in three key complementary areas: organizing background
knowledge, capturing local knowledge, and disseminating contextualized knowledge
usefully, appropriately, and on time.1Rosenberg, M. J. (2001). E-learning: Strategies
for delivering knowledge in the digital age. New York: McGraw-Hill

200
Bio-inspired Restoration of Aged Con-
crete Edifices (BRACE)
Dr. Matthew Pava

https://www.darpa.mil/program/bio-inspired-restoration-of-aged-concrete-edifices.html

GPT-3 Summary: The BRACE program will develop bio-inspired approaches to

repair and prolong the serviceability of existing concrete structures and pavements.
These approaches will 1) integrate deep within aged concrete to form a healing “vas-
culature” for ongoing damage repair; and 2) combine with new concrete to increase
the durability of runway patch repairs.

Full Details: The Bio-inspired Restoration of Aged Concrete Edifices (BRACE) pro-
gram aims to prolong the serviceability of Department of Defense (DoD) structures
and airfield pavements by integrating a self-repair capability into existing concrete.
The DoD relies on steel-reinforced concrete structures such as missile silos and naval
piers that are many decades old, not easily replaced, and subject to cracking and corro-
sive deterioration. The DoD also relies on concrete airfield pavements in expeditionary
settings, which are vulnerable to damage from overuse or attack and require rapid
repair under challenging circumstances. Unfortunately, state-of-the-art approaches to
maintain concrete are one-time interventions, limited to remediation of defects at or
near the surface, and typically necessitate down-time for critical assets. No current
technology provides ongoing crack repair and prevention for defects deep inside ex-
isting aged concrete or prolonged repair of damaged airfield pavements. Inspired by
vascular systems that support repair in multicellular organisms and ecosystems, the
BRACE program will develop bio-inspired approaches that 1) integrate deep within
aged concrete to form a healing “vasculature” for ongoing damage repair; and 2)
combine with new concrete to increase the durability of runway patch repairs.

201
PREventing EMerging Pathogenic
Threats (PREEMPT)
Dr. Kristen Jordan

https://www.darpa.mil/program/preventing-emerging-pathogenic-threats.html

GPT-3 Summary: GOAL: The PREEMPT program seeks to prevent the emer-
gence of human-pathogenic threats by targeting viral pathogens within animal reser-
voirs and insect vectors.
METHODS: The program combines biosurveillance and modeling with novel tech-
nologies for treating or containing high-risk pathogens at their source.
IMPACT: By focusing on animal reservoirs and insect vectors, the program seeks to
identify opportunities for intervention that exploit the evolutionary bottlenecks and
transmission factors that enable pathogen species jump.

Full Details: United States military forces deploy to remote locations around the
world, often in areas where emerging infectious diseases are common. The PREvent-
ing EMerging Pathogenic Threats (PREEMPT) program seeks to preserve military
readiness by protecting against the infectious disease threat; however, rather than
treating people, PREEMPT targets viral pathogens within the animal reservoirs and
insect vectors where many diseases originate before they spill over into humans. The
program combines biosurveillance and modeling with novel technologies for treat-
ing or containing high-risk pathogens at their source to prevent the emergence and
reemergence of human-pathogenic threats. PREEMPT builds on recent advances in
understanding of host-pathogen genetic interactions and mechanisms of adaptation
across species, emerging analytic tools to predict what species might carry poten-
tial human-pathogens, and novel capabilities to predict geographic “hot spots” where
an animal-to-human viral jump is likely. By focusing far-forward biosurveillance on
animal reservoirs and insect vectors, and applying high-throughput molecular tech-
nologies and next-generation sequencing to determine how viruses evolve within a
species, PREEMPT seeks to identify opportunities for intervention that exploit the
evolutionary bottlenecks and transmission factors that enable pathogen species jump.
Although PREEMPT pursues technology for eventual deployment, research is per-
formed entirely in controlled laboratory facilities, including planned proof-of-concept
demonstrations at the end of the program. If the program is successful, potential fu-
ture field trials would take place under the auspices of other government stakeholders
following all standard protocols for biosafety.

202
Epigenetic CHaracterization and Ob-
servation (ECHO)
CDR Jean-Paul Chretien

https://www.darpa.mil/program/epigenetic-characterization-and-observation.html

GPT-3 Summary: The ECHO program’s goal is to develop a portable device that
can read an individual’s epigenome in order to determine if they have been exposed to
weapons of mass destruction, their precursors, or infectious diseases. The technology
would work by quickly reading an individual’s epigenome from a biological sample and
conducting analysis to reveal possible exposure events. The ECHO system would offer
an enormous advantage over current forensic and diagnostic screening technologies as
it will identify the type of exposure and when it occurred without the need for trace
amounts of the threat in the sample.

Full Details: The Epigenetic CHaracterization and Observation (ECHO) program

aims to diminish the threat posed by weapons of mass destruction (WMD). To do
this, the program is building a man-portable device that analyzes an individual’s
epigenetic “fingerprint” to potentially reveal a detailed history of that individual’s
exposure to WMD or their precursors. DARPA envisions that the same technology
could provide rapid diagnostics for troops who may have been exposed to threat
agents or who may be suffering from infections, providing a timely signal to apply
effective medical countermeasures.ECHO technology would work by quickly reading
an individual’s epigenome — a part of human biology that helps our body respond to
the constantly changing world around us — and identifying the epigenetic signatures
indicative of WMD, precursor, or infectious disease exposure. Though a person’s un-
derlying genetic code is stable, life events can leave discernible marks on the genome
that modify how genes are expressed. The epigenome is the combination of all such
natural modifications, and while epigenetic modifications can register within seconds
to minutes, they can potentially imprint the genome for decades, leaving a unique,
time-stamped biography of an individual’s exposures. The ECHO system would read
someone’s epigenome from a biological sample such as a finger prick or nasal swab,
and conduct analysis to reveal possible exposure events even when other physical ev-
idence has been erased. ECHO technology would offer an enormous advantage over
state-of-the-art forensic and diagnostic screening technologies as it will identify the
type of exposure and when it occurred — all without the need for trace amounts of the
threat in the sample. Program performers will develop new approaches to identifying
and characterizing epigenetic signatures from WMD, precursor, or infectious disease

203
exposure events, and create new bioinformatics tools to perform forensic analyses
with high sensitivity, specificity, and temporal resolution. They will have to create a
device capable of performing multiple molecular analyses and onboard bioinformatics
in 30 minutes or less, compared to an average of two days using current laboratory
processes. By the end of the program, DARPA’s plans to demonstrate ECHO capa-
bilities in a man-portable device that could be used by an operator in austere settings
with only minimal training.

204
Aerial Dragnet
Dr. Paul Zablocky

https://www.darpa.mil/program/aerial-dragnet.html

GPT-3 Summary: GOAL: The Aerial Dragnet program will aim to detect and
track small unmanned aerial systems (UAS) in urban terrain.

TECHNOLOGY METHODS: The program will seek innovative technologies

to provide persistent, wide-area surveillance of all UAS operating below 1,000 feet in
a large city.

IMPACT: The system could find civilian application to help protect U.S. metropolitan
areas from UAS-enabled terrorist threats.

Full Details: Airspace for the flying public today is perpetually congested yet re-
markably safe, thanks in no small part to a well-established air traffic control system
that tracks, guides and continuously monitors thousands of flights a day. When it
comes to small unmanned aerial systems (UAS) such as commercial quadcopters,
however, no such comprehensive tracking system exists. And as off-the-shelf UAS
become less expensive, easier to fly, and more adaptable for terrorist or military pur-
poses, U.S. forces will increasingly be challenged by the need to quickly detect and
identify such craft—especially in urban areas, where sight lines are limited and many
objects may be moving at similar speeds. DARPA’s Aerial Dragnet program aims to
achieve the technically difficult goal of detecting and tracking small UAS in urban
terrain. The program seeks innovative technologies to provide persistent, wide-area
surveillance of all UAS operating below 1,000 feet in a large city. While Aerial Drag-
net’s focus is on protecting military troops operating in urban settings overseas, the
system could ultimately find civilian application to help protect U.S. metropolitan
areas from UAS-enabled terrorist threats.

205
Consortium for Execution of Ren-
dezvous and Servicing Operations (CON-
FERS)
Ms. Ana Saplan

https://www.darpa.mil/program/consortium-for-execution-of-rendezvous-and-servicing-
operations.html

GPT-3 Summary: The goal of the CONFERS program is to create an indus-

try/government forum where experts can collaborate and develop standards for on-
orbit servicing operations. The standards would be used by commercial on-orbit ser-
vicing providers to ensure best practices for operational safety. The program would
be transferred to industry by 2021.

Full Details: Recent technological advances have made the longstanding dream
of on-orbit robotic servicing of satellites a near-term possibility. The potential ad-
vantages of that unprecedented capability are enormous. Instead of designing their
satellites to accommodate the harsh reality that, once launched, their investments
could never be repaired or upgraded, satellite owners could use robotic vehicles to
physically inspect, assist, and modify their on-orbit assets. That could significantly
lower construction and deployment costs while dramatically extending satellite utility,
resilience, and reliability. In fact, efforts to achieve the goal of on-orbit servicing are
already underway, including DARPA’s Robotic Servicing of Geosynchronous Satellite
(RSGS) program, which focuses on services for satellites in geosynchronous orbit. But
these efforts all face a major roadblock: the lack of clear, widely accepted technical
and safety standards for responsible performance of on-orbit activities involving com-
mercial satellites, including rendezvous and proximity operations (RPO) that don’t
involve physical contact with satellites and robotic servicing operations that would.
Without these standards, the long-term sustainability of outer space operations is
potentially at risk.DARPA’s Consortium for Execution of Rendezvous and Servic-
ing Operations (CONFERS) program aims to help overcome these challenges and
provide the foundation for a new commercial repertoire of robust space-based capa-
bilities. CONFERS envisions a permanent, self-sustaining, and independent forum
where industry could collaborate and engage with the U.S. Government about on-
orbit servicing. This industry/government forum would be composed of experts from
throughout the space community. DARPA, primarily in partnership with NASA, will
bring decades of operational experience from government missions to the consortium.
Participants would leverage best practices from government and industry to research,

206
develop, and publish non-binding, consensus-derived technical and safety standards
that servicing providers and clients for on-orbit servicing operations would adopt.
In doing so, the program would provide a clear technical basis for definitions and
expectations of responsible behavior in outer space. The standards would be broad
enough to allow individual companies to pursue their own implementations of these
standards to suit their individual businesses, while assuring that the implementations
adhere to best practices for operational safety.DARPA aims to transfer CONFERS
leadership and funding to industry by 2021, when the Agency is scheduled to perform
its first on-orbit demonstration of RSGS. The Agency also intends by that time to
release the standards for general use by commercial on-orbit servicing organizations.

207
Software Defined Hardware (SDH)
Dr. Howard Shrobe

https://www.darpa.mil/program/software-defined-hardware.html

GPT-3 Summary: The goal of the Software Defined Hardware (SDH) program is
to build runtime-reconfigurable hardware and software that enables near ASIC per-
formance without sacrificing programmability for data-intensive algorithms. The pro-
gram will target very fast hardware reconfiguration speeds and dynamic compilation
in order to create systems that can take advantage of data-dependent optimizations
that even today’s ASICs cannot exploit. If successful, SDH will be able to provide
the Department of Defense with a low-cost way to develop and run data-intensive,
data-exploitation algorithms.

Full Details: In modern warfare, decisions are driven by information. That infor-
mation can come in the form of thousands of sensors providing information, surveil-
lance, and reconnaissance (ISR) data; logistics/supply-chain and personnel perfor-
mance measurements; or a host of other sources and formats. The ability to exploit
this data to understand and predict the world around us is an asymmetric advantage
for the Department of Defense (DoD).Utilizing this data relies on computational al-
gorithms running at a huge scale. Today, developers are limited in their ability to
run these algorithms efficiently because they generally have to trade the efficiency of
their algorithms with that of the available hardware architecture implementations.
To combat this challenge, one solution is to design and fabricate application specific
integrated circuits (ASICs) — customized hardware designed to maximize the run-
time efficiency of a specific algorithm. However, ASICs typically cost hundreds of
millions of dollars and take many years to develop. Once developed, they can per-
form exactly one class of computation because they were designed and specialized
for specific application tasks. Because these systems are so specifically tailored and
costly, their creation is often limited to the highest priority algorithms. For prob-
lems that cannot afford this level of investment, compute efficiency is sacrificed by
implementing solutions such as software on general-purpose processors or field pro-
grammable gate arrays (FPGAs). Often, this results in application implementations
that are thousands of times worse than optimal. The goal of the Software Define
Hardware (SDH) program is to build runtime-reconfigurable hardware and software
that enables near ASIC performance without sacrificing programmability for data-
intensive algorithms. Under the program, data-intensive algorithms are defined as
machine learning and data science algorithms that process large volumes of data and
are characterized by their usage of intense linear algebra, graph search operations,

208
and their associated data-transformation operators. The SDH program aims to cre-
ate hardware/software systems that allow data-intensive algorithms to run at near
ASIC efficiency without the cost, development time, or single application limitations
associated with ASICs. If successful, SDH will result in the ability to develop and run
data-intensive, data-exploitation algorithms at very low cost, and, consequently, en-
able pervasive use of big-data solutions for a wide range of DoD applications including
ISR, predictive logistics, decision support, and beyond. SDH Program Concept and
Structure The SDH program will create malleable hardware/software architectures
that, unlike ASICs, allow an application to defer hardware configuration to runtime.
SDH will enable 1) the optimization of code and hardware dynamically when input
data change, and 2) the reuse of hardware for new problems and new algorithms to
solve existing problems. In order to accomplish these objectives, SDH targets very fast
hardware reconfiguration speeds and dynamic compilation. If successful, SDH will
be able to take advantage of data-dependent optimizations that even today’s ASICs
cannot exploit.The SDH program will measure compute efficiency (in terms of giga-
operations per watt (GOPs/Watt)). By the end of the program, it is expected that
SDH systems to be at efficiencies within 5X of ASICs and 500-1000X better than
CPU implementations with the same programmability as current NumPy/Python
implementations.

209
The Waveform Agile Radio-frequency
Directed ENergy (WARDEN)
Dr. David K. Abe

https://www.darpa.mil/program/waveform-agile-radio-frequency-directed-energy.html

GPT-3 Summary: The Waveform Agile Radio-frequency Directed ENergy

(WARDEN) program seeks to develop hardware, theory, and computational models
to extend the range and effectiveness of high-power microwave (HPM) systems for
backdoor attacks.

The program comprises three technical areas: HPM Traveling-wave Amplifier

development, Rapid Assessment and Numerical Generation of EM Response
(RANGER), and Agile Waveform Development. These key areas will be central in
developing the necessary hardware, modeling tools and waveforms need to make
WARDEN successful.

Full Details: The Waveform Agile Radio-frequency Directed ENergy (WARDEN)

program seeks to develop hardware, theory, and computational models to extend
the range and effectiveness of high-power microwave (HPM) systems for backdoor
attacks. HPM systems are a class of directed energy weapons (DEWs) that use
electromagnetic (EM) radiation to disrupt, disable, or damage targeted electronic
components and circuits. Electromagnetic radiation can couple into targets in-band
via intentional ports such as antennas (“front door”) or via unintentional coupling
paths such as seams, apertures, and cable entry points (“back door”). The advantages
of HPM systems include non-kinetic, wide-area effects at large stand-off distances;
deep magazines; operation in adverse environmental conditions; and speed-of-light
engagement. In general, current HPM systems use oscillators as their RF source.
These systems operate at a fixed frequency, are not readily tunable, and lack the
phase coherence necessary for power combining. Front-door systems have the longest
range, but their effectiveness is limited to the specific classes of targets for which they
were designed. Back-door systems are effective against a wider variety of targets, but
their range is limited by EM coupling inefficiencies arising from their lack of frequency
tunability. Agile waveforms, combined with broadband high-power amplifiers, can
reduce the threshold of the susceptibility of targeted systems to backdoor attacks
and significantly extend the range and effectiveness of the HPM weapon system.
WARDEN is developing flexible HPM technology that can be useful against a wide
variety of target types.The WARDEN program comprises three technical areas: HPM

210
Traveling-wave Amplifier development, Rapid Assessment and Numerical Generation
of EM Response (RANGER), and Agile Waveform Development. These key areas
will be central in developing the necessary hardware, modeling tools and waveforms
need to make WARDEN successful.

211
Adapting Cross-Domain Kill-Webs
(ACK)
Dr. Greg Kuperman

https://www.darpa.mil/program/adapting-cross-domain-kill-webs.html

GPT-3 Summary: The goal of the Adapting Cross-Domain Kill-Webs (ACK) pro-
gram is to provide a decision aid for mission commanders to help them rapidly identify
and select options for tasking assets within and across organizational boundaries. The
program is leveraging commercially developed solutions adapted from e-commerce and
is modifying them to help address these challenges. These algorithms are then com-
bined with a “Virtual Liaison” service abstraction from a much larger “Capability
Marketplace” to enable decentralized Joint All-Domain adaptation. If the program
is successful, the technology developed under ACK will be an important enabler for
the

Full Details: The goal of the Adapting Cross-Domain Kill-Webs (ACK) program is
to provide a decision aid for mission commanders to assist them with rapidly iden-
tifying and selecting options for tasking – and retasking – assets within and across
organizational boundaries. Specifically, ACK will assist users with selecting sensors,
effectors, and support elements across military domains (space, air, land, surface,
subsurface, and cyber) that span the different military Services to deliver desired
effects on targets. Instead of limited, monolithic, pre-defined kill chains, these more
disaggregated forces can be used to formulate adaptive “kill webs” based on all of
the options available. There are three major challenges to realizing the ACK goals.
First, during execution, mission commanders have little or no insight into what ca-
pabilities are available across domains and what capacity or quality of service other
domains may be able to offer. Second, each domain has its own set of authorities
and missions they are tasked to service, making it challenging to assess meaningful
tradeoffs of the “value” or “cost” of supporting new missions originating from an-
other domain versus their own current set of missions. Third, given a set of diverse
cross-domain kill web options, decision-makers need a way to compare them rapidly
and select the “best” option. The ACK program is leveraging commercially devel-
oped solutions adapted from e-commerce and is modifying them to help address these
challenges. These algorithms are then combined with a “Virtual Liaison” service ab-
straction from a much larger “Capability Marketplace” to enable decentralized Joint
All-Domain adaptation. The program is currently investigating the right bid and
offer language necessary between consumers and suppliers in the marketplace along

212
with evaluating the relative merits of the different technical approaches. These in-
clude assessing the processing and bandwidth requirements along with characterizing
any limitations with respect to marketplace size (relative and absolute numbers of
consumers, suppliers, constraints, etc), Virtual Liaison architectural representation,
and planning horizon. ACK will enable multiple warfighters to define distributed
effects and adapt them at combat speed using a shared set of resources. This will
increase lethality and capitalize on latent capacity within the existing warfighting
force structure by pairing the right sensor and weapon together for a given target
and operational problem. This will also create greater resilience by enabling rapid
substitutions if a capability is lost and will make the current force more efficient by
enabling better sharing of resources across domains to balance tasking loads.Finally,
the tools developed in ACK will aid with the protection of sensitive capabilities by
allowing capability providers (suppliers) to offer assets across domains in terms of the
effects they can provide, without exposing details regarding how those effects will be
achieved (i.e. without revealing sources and methods). If the program is successful,
the technology developed under ACK will be an important enabler for the new Joint
All-Domain concepts that the Services are pursuing at both operational and tactical
levels.

213
Focused Pharma
Dr. Matthew Pava

https://www.darpa.mil/program/focused-pharma.html

GPT-3 Summary: GOAL: Develop novel drugs with broad-spectrum therapeutic

activity and rapid onset of effect for treating a range of neuropsychiatric conditions
of high need within the Department of Defense.

TECHNOLOGY METHODS: Research in medicinal chemistry and neurophar-

macology to develop novel drugs that specifically engage select neurotransmitter
receptor subtypes and demonstrate downstream signaling pathway bias.

IMPACT: Generate and validate completely novel chemical compounds that

show both increased effectiveness and reduced toxic side effects relative to current
standards of treatment.

Full Details: Military service members experience an increased lifetime risk of neu-
ropsychiatric conditions, such as depression, post-traumatic stress, and substance
abuse. These conditions are substantially more prevalent in both the active duty
population and veterans relative to civilians. Current treatment options for neu-
ropsychiatric conditions typically involve some combination of psychotherapy, electri-
cal stimulation, and/or psychopharmacology. Each of these strategies has significant
limitations that hinder use, notably effectiveness, time-to-therapeutic onset, and/or
invasiveness. There is a pressing need for fast-acting and safe therapeutics that can be
administered acutely to mitigate the deleterious consequences of neuropsychiatric ill-
ness. The Focused Pharma program seeks to develop novel drugs with broad-spectrum
therapeutic activity and rapid onset of effect for treating a range of neuropsychiatric
conditions of high need within the Department of Defense. The program supports
research in medicinal chemistry and neuropharmacology to develop novel drugs that
specifically engage select neurotransmitter receptor subtypes and demonstrate down-
stream signaling pathway bias. The terminal goal of the program is to generate and
validate completely novel chemical compounds that show both increased effectiveness
and reduced toxic side effects relative to current standards of treatment.

214
Harnessing Enzymatic Activity for
Lifesaving Remedies (HEALR)
Dr. Seth Cohen

https://www.darpa.mil/program/healr.html

GPT-3 Summary: GOAL: Develop new medical countermeasures (MCMs) to treat

microbial infections by harnessing cellular processes.
TECHNOLOGY METHODS: Develop tools to target microbial pathogens for protein
degradation, new protein degradation modalities, and a platform that leverages these
advances.
IMPACT: Create new ways to treat microbial infections that are more flexible and
responsive to emerging threats.

Full Details: Microbial infections are a problem of particular concern to the Depart-
ment of Defense (DoD). The DoD has long recognized the warfighter’s outsized risk
of exposure to infectious disease, including the rise of antimicrobial resistant (AMR)
and multidrug resistant (MDR) pathogens that have challenged military wound care
in Iraq and Afghanistan. Furthermore, the responsibility of the DoD to protect
the homeland encompasses biological threat agents, including bacterial threats, for
which effective countermeasures are critical.HEALR seeks to establish an orthogo-
nal approach to treating microbial infections by harnessing advancements in recruit-
ing native cellular host machinery to recognize and eliminate disease-related targets.
Specifically, HEALR will develop new medical countermeasures (MCMs) that result
in host-driven degradation or deactivation of pathogen targets. By harnessing innate
cellular processes, approaches such as proteolysis targeting chimeras (PROTACs)
and similar methods can achieve superior outcomes over existing therapies.HEALR
aims to develop: (1) tools to target microbial pathogens for protein degradation (i.e.,
target-binding ligand development), (2) new protein degradation modalities to en-
able new pathways to protein degradation or deactivation (i.e., ‘NewTACs’), and (3)
a platform that leverages these advances to permit flexible and rapid response to
emerging threats.

215
Technologies for Mixed-mode Ultra
Scaled Integrated Circuits (T-MUSIC)
Dr. James Wilson

https://www.darpa.mil/program/technologies-for-mixed-mode-ultra-scaled-integrated-
circuits.html

GPT-3 Summary: GOAL: Develop new technologies for mixed-mode integrated

circuits (ICs) to improve performance in a variety of applications.
METHODS: Develop advanced materials, device processing, and innovative mixed-
mode circuit designs based on an advanced CMOS fabrication platform.
IMPACT: Improve speed and accuracy of integrated mixed-mode electronics, enabling
U.S. technology leadership in telecommunications, automotive radars, and edge arti-
ficial intelligence sensors.

Full Details: High performance mixed-mode electronics provide the interface be-
tween analog signals in the physical world and digital signals in the information do-
main. The performance of mixed-mode interfaces directly affects applications ranging
from the national telecommunications infrastructure to defense electromagnetic sys-
tems for sensing, communications, and electronic warfare. Current implementations
of these interfaces in traditional complementary metal-oxide-semiconductor (CMOS)
technologies are increasingly constrained by the limits of MOS transistor scaling.The
Technologies for Mixed-mode Ultra Scaled Integrated Circuits (T-MUSIC) program
aims to develop advanced materials, device processing, and innovative mixed-mode
circuit designs based on an advanced CMOS fabrication platform to vastly improve
the speed and accuracy of integrated mixed-mode electronics. Specifically, T-MUSIC
aims to develop advanced on-shore mixed-mode foundry technologies monolithically
combining cutting-edge CMOS with analog transistors capable of exceeding 600 GHz.
The program also aims to demonstrate innovative broadband precision mixed-mode
integrated circuits leveraging these foundry technologies.T-MUSIC will enable broad
access to mixed-mode technology in on-shore foundries to support U.S. commercial
and defense technology leadership in dual-use applications including telecommunica-
tions, automotive radars, and edge artificial intelligence sensors.

216
Pandemic Prevention Platform (P3)
Dr. Amy Jenkins

https://www.darpa.mil/program/pandemic-prevention-platform.html

GPT-3 Summary: The goal of the P3 program is to develop a platform that can
rapidly produce medical countermeasures against any known or unknown infectious
disease threat. The methods used will involve developing new ways to grow viruses for
testing, identify and mature protective antibodies, and deliver nucleic acid constructs
into patients. The impact of this program will be the ability to contain outbreaks of
disease by rapidly producing countermeasures.

Full Details: The Pandemic Prevention Platform (P3) program aims to support mil-
itary readiness and global stability through pursuit of novel methods to dramatically
accelerate discovery, integration, pre-clinical testing, and manufacturing of medical
countermeasures against infectious diseases. P3 confronts the reality that Depart-
ment of Defense (DoD) personnel are not only deployed around the world for routine
operations, but are often among the first responders to outbreaks of emerging or
re-emerging disease with pandemic potential (e.g., Ebola). P3 aims specifically to
develop a scalable, adaptable, rapid response platform capable of producing rele-
vant numbers of doses against any known or previously unknown infectious threat
within 60 days of identification of such a threat in order to keep the outbreak from
escalating and decrease disruptions to the military and homeland. State-of-the-art
medical countermeasures often take many months or even years to develop, produce,
distribute, and administer. The envisioned P3 platform would cut response time
to weeks and stay within the window of relevance for containing an outbreak. P3
focuses on rapid discovery, characterization, production, testing, and delivery of effi-
cacious DNA- and RNA-encoded medical countermeasures, a foundational technology
pioneered by DARPA under the Autonomous Diagnostics to Enable Prevention and
Therapeutics (ADEPT) program that provides the body with instructions on how to
immediately begin producing protective antibodies against a given threat. The P3
program seeks to unlock the potential of these coded genetic constructs—establishing
them as the basis for a threat-agnostic platform technology—by achieving and inte-
grating breakthroughs in three key areas: novel approaches for the growth of viruses
for use in testing and evaluation of countermeasures; rapid identification and matura-
tion of protective antibodies to increase their potency; and novel technologies for the
delivery of nucleic acid constructs into patients to encode the antibody of interest and
produce a protective response. A principal benefit of the nucleic-acid-based approach
to limiting the spread of infection is that genetic constructs introduced into the body

217
would process quickly and not integrate into an individual’s genome. Similarly, the
antibodies produced in response to treatment would only be present in the body for
weeks to months. This is consistent with DARPA’s intent to safely deliver transient
immunity, halting the spread of disease by creating a firewall, and buying time for
longer-term medical responses to be developed and deployed.

218
Military Tactical Means (MTM)
Mr. Chris Simi

https://www.darpa.mil/program/military-tactical-means.html

GPT-3 Summary: GOAL: Develop sensors and exploitation techniques to detect

high-value targets.

METHODS: Use small platforms in wide-area searches. Identify targets and

maintain positive chain-of-custody hand-offs.

IMPACT: Improve targeting capabilities for the military.

Full Details: The MTM program aims to develop sensors and exploitation tech-
niques capable of performing wide-area searches to detect high-value targets. Pro-
gram design will provide the ability to identify high-value adversary targets and to
maintain positive chain-of-custody hand-offs. In addition, MTM will have the poten-
tial to be used in highly proliferated systems, such as small platforms.

219
Dynamic Network Adaptation for Mis-
sion Optimization (DyNAMO)
Mr. Aaron Kofford

https://www.darpa.mil/program/dynamic-network-adaptation-for-mission-optimization.html

GPT-3 Summary: GOAL: The DyNAMO program is developing and testing

technologies to enable adaptive, mission-responsive networking among diverse
platforms in contested environments.

TECHNOLOGY METHODS: The program is addressing optimization within

legacy and future military networks, interactions between networks, and availability
of necessary network services to support mission success.

IMPACT: The DyNAMO program will conduct experiments with the new
technology on various radio platforms to show interoperability in field events with
Navy and Marine units.

Full Details: The Dynamic Network Adaptation for Mission Optimization (Dy-
NAMO) program is developing and testing technologies that enable adaptive, mission-
responsive networking among diverse platforms in contested environments. Wire-
less networks have evolved into complex systems having many configurable parame-
ters/features, including link data rates, power settings, inter-network gateways, and
security associations. The optimal settings for these features vary greatly depending
on the mission for which the network is deployed and the environment in which it is
operating. Currently, the majority of these features are optimized off-line for specific
scenarios and assumptions and are pre-set before use in a mission. There is no capa-
bility for the settings to adapt if the actual mission or environment differs from the
original assumptions used to configure the network. The problem is exacerbated in
scenarios in which intelligent adversaries can affect the topology and operation of the
network unpredictably and on short timescales. Furthermore, future operations will
include multiple, different radios interconnected on the same platform, which requires
adaptation of the interaction among different networks. The DARPA DyNAMO pro-
gram is developing information-centric approaches to bridge disparate networks and
to adaptively configure and control networks and networks of networks for operation
in dynamic and contested environments. The program is addressing optimization
within legacy and future military networks, interactions between networks, and avail-
ability of necessary network services to support mission success. During the summer

220
and autumn of 2020, the DyNAMO program will conduct experiments with the new
technology on various radio platforms to show interoperability in field events with
Navy and Marine units.

221
Reduction of Entropy for Probabilis-
tic Organization (REPO)
Mr. Daniel Ridge

https://www.darpa.mil/program/reduction-of-entropy-for-probabilistic-organization.html

GPT-3 Summary: GOAL: The REPO program will develop techniques to reduce
the risk of obsolescence management and re-engineering of legacy designs.

TECHNOLOGY METHODS: Conduct initial flow chart analysis and link block
diagram data to text documentation in a single cohesive structure. Automate
analysis of ECAD and MCAD to identify and export notable features in graphs
and human-readable summary format for analysts. Generate chord diagram models
showing relationships between dataset entities and permit an analyst to consider
relative magnitudes in comparison to alternative arcs to understand previously
undiscoverable connections.

IMPACT: The REPO program

Full Details: Design and development programs in aerospace and defense often fea-
ture high complexity and long design times, which translates into long sustainment
times to pay off the incurred development costs. When the inevitable technical refresh
effort for such designs begins, the design intent for the equipment in question can of-
ten be problematic to divine as the original engineers and personnel are no longer
available (or possess the requisite memories).The Reduction of Entropy for Proba-
bilistic Organization (REPO) AI Exploration (AIE) effort aims to develop techniques
to enable analysts to process large, unstructured, and unorganized documentation
datasets that contain electrical Computer Aided Design (ECAD) schematic files, me-
chanical CAD (MCAD) drawings, and flowchart/block diagram data. The goal is to
implement REPO techniques into tools that reduce the technical and program risk
of obsolescence management and re-engineering of legacy designs.Specific capabilities
sought by REPO include:Conduct initial flow chart analysis and link block diagram
data to text documentation in a single cohesive structureAutomate analysis of ECAD
and MCAD to identify and export notable features in graphs and human-readable
summary format for analystsGenerate chord diagram models showing relationships
between dataset entities and permit an analyst to consider relative magnitudes in
comparison to alternative arcs to understand previously undiscoverable connections-
For more information, see the REPO Program Announcement.

222
Cornerstone
Dr. Matthew Pava

https://www.darpa.mil/program/cornerstone.html

GPT-3 Summary: GOAL: Develop safe and effective countermeasures to prevent

brain injury.
TECHNOLOGY METHODS: Develop prophylactic therapeutics that prevent injury
by blocking the initiation of harmful signaling.
IMPACT: Improve mission execution and warfighter protection by addressing the gap
in treatments for TBI.

Full Details: The Cornerstone program aims to develop safe and effective Inte-
grated Countermeasures to prevent brain injury. Military personnel face a high risk
of traumatic brain injury (TBI) resulting in a debilitating long-term burden both
to the person, as well as the healthcare infrastructure. At this time, no treatments
for TBI have progressed past Phase II clinical trials. Cornerstone aims to improve
mission execution, and warfighter protection by addressing this gap. Ultimately, the
development of prophylactic therapeutics that prevent injury by blocking the initia-
tion of harmful signaling will sustain warfighter readiness, and facilitate a strategic,
operational, and tactical advantage.

223
Habitus
Dr. Bartlett Russell

https://www.darpa.mil/program/habitus.html

GPT-3 Summary: GOAL: Habitus will allow military operators to have an insider’s
view of local knowledge in order to make better decisions.
METHODS: New engagement mechanisms and methodologies for creating predictive,
causal models of local systems.
IMPACT: A generalizable, adaptive, and self-sustaining capability across regions and
populations.

Full Details: What is opaque to outsiders is often obvious – even if implicit –

to locals. Habitus aims to capture and make local knowledge available to military
operators, providing them with an insider view to support decision making. This in-
volves developing: 1) new engagement mechanisms; and 2) methodologies for creating
predictive, causal models of local systems based on input from locals. The resulting
capability would be generalizable across regions and populations, adaptive as societies
change over time, and self-sustaining for maintainability and persistence.

224
Reefense
Dr. Catherine Campbell

https://www.darpa.mil/program/reefense.html

GPT-3 Summary: Reefense will develop structures that can rapidly be deployed
to provide protection and then facilitate the growth of calcareous organisms. This
will happen through recent innovations in materials science, hydrodynamic modeling,
adaptive biology, and multiple effectors.

Full Details: Reefense seeks to develop self-healing, hybrid biological and engi-
neered reef-mimicking structures to mitigate the coastal flooding, erosion and storm
damage that increasingly threaten civilian and DoD infrastructure and personnel.
Under Reefense, custom wave-attenuating base structures will promote calcareous
reef organism (coral or oyster) settlement and growth, which will enable the system
to self-heal and keep pace with sea level rise over time. A system will be put in
place that will also attract non-reef building organisms necessary to help maintain a
healthy, growing system. Finally, adaptive biology (other than genetically modified
organisms) will enable improved coral and oyster resilience against disease and tem-
perature stress, to ensure compatibility with a changing environment.The Reefense
strategy includes employing recent innovations in materials science, hydrodynamic
modeling, adaptive biology, and multiple effectors to develop growing structures that
are optimized to rapidly implement coastal defenses suited to a changing environment.
By co-designing structures with biology such that they can be rapidly deployed to
provide immediate protection and then persistently facilitate the growth of calcareous
organisms, the protective Reefense structures will be rapidly enhanced by what would
normally require decades of biological structure-building, but which can be achieved
in a matter of months to years.

225
Learning Introspective Control (LINC)
Mr. John-Francis Mergen

https://www.darpa.mil/program/learning-introspective-control.html

GPT-3 Summary: GOAL: Develop technologies that enable physical systems to

adapt to unforeseen events in real-time.
TECHNOLOGY METHODS: Develop machine learning-based introspection tech-
nologies that continually compare the real-time behavior of a military platform with
a learned model of the platform.
IMPACT: Allow operators to maintain effective control and trust in the platform.

Full Details: Control design currently aims to model the range of operating envi-
ronments that are anticipated at design time. Plans can fail, however, when physical
attacks, unforeseen conditions, or unanticipated use places the system outside the de-
sign envelope. Custom tailored aftermarket remedies are not commonly available and
require a skilled technician to install, and modifications to existing systems through
procurement channels can take months or years. The Department of Defense (DoD)
systems are particularly long-lived, so ongoing adaptation would permit continual
modification as missions and theaters change, providing a strategic advantage over
an adversary.The Learning Introspective Control (LINC) program aims to develop
machine learning-based introspection technologies that enable physical systems, with
specific interest in ground vehicles, ships, drone swarms, and robotic systems, to
respond to events not predicted at design time. LINC technologies endeavors to up-
date control laws as required in real time while providing guidance and situational
awareness to the operator, whether that operator is human or an autonomous con-
troller.The current approach to handling platform damage or modification places the
burden of recovery and control on the operator. LINC seeks to develop machine intro-
spection and learning technologies that continually compare the real-time behavior of
a military platform – as measured by on-board sensors – with a learned model of the
platform, determine whether the observed behavior differs from that model in ways
that might compromise stability and control, and implement an updated control law
when required and effectively communicate those updates to the operator. This al-
lows the operator to maintain effective control and trust in the platform. Additional
information is available in the LINC BAA.

226
Digital RF Battlespace Emulator (DRBE)
Mr. John Davies

https://www.darpa.mil/program/digital-rf-battlespace-emulator.html

GPT-3 Summary: The goal of the DRBE program is to create a virtual radio
frequency (RF) environment for developing, training, and testing advanced RF sys-
tems. The DRBE will use novel computing architectures to generate computational
performance in the double-digit petaFLOPs class while maintaining single-digit mi-
crosecond scale end-to-end latency. This will allow for the creation of a large-scale
virtual RF test range.

Full Details: The Digital RF Battlespace Emulator (DRBE) program aims to create
the world’s first, large-scale, virtual radio frequency (RF) environment for develop-
ing, training, and testing advanced RF systems, such as radar and electronic warfare
(EW) systems. The target DRBE environment will enable numerous RF systems to
interact with each other in a fully closed-loop RF arena, replicating dense, respon-
sive, real-world RF environments. RF systems increasingly use artificial intelligence
(AI) to help automate and augment capabilities for defense use. To help address the
rigorous demands of testing and training these AI-enabled systems 24/7/365, virtual
simulators are needed. In other domains, such as training modern fighter aircraft pi-
lots, simulators are already in use to augment real-world aircraft flight hours. Current
simulated environments, however, rely on conventional computing that is incapable of
generating the computational throughput and speed to accurately replicate real-world
RF interactions, model the scale of physical test ranges, or meet the technical require-
ments of more complex systems. DRBE is exploring novel computing architectures to
enable the creation of a new breed of High Performance Computing (HPC) – dubbed
“Real Time HPC” or RT-HPC. The goal of the RT-HPC is to generate computational
performance in the double-digit petaFLOPs class while maintaining single-digit mi-
crosecond scale end-to-end latency. By balancing computational throughput with
extreme low latency, DRBE should be capable of generating high-fidelity RF envi-
ronments. To support the creation of RT-HPC and the virtual RF test range, DRBE
will: 1) develop novel processors and application-specific integrated circuits (ASICs)
that realize the low-latency, high compute capacity vision; 2) assemble these ASICs
into a multi-processor RT-HPC system; and 3) design and integrate the necessary
tools to demonstrate the use of the RT-HPC as a large-scale virtual RF test range.

227
Detect It with Gene Editing Tech-
nologies (DIGET)
CDR Jean-Paul Chretien

https://www.darpa.mil/program/detect-it-with-gene-editing-technologies.html

GPT-3 Summary: The goal of the DIGET program is to develop field-forward diag-
nostic and biosurveillance technologies that enable detection of any threat, anytime,
anywhere. This will be achieved by designing, developing, prototyping, and deploying
two novel nucleic acid detection devices: 1) a disposable point-of-need diagnostic for
up to 10 pathogens and host biomarkers, and 2) a massively multiplexed detection
(MMD) device for 1,000 or more targets. If successful, the disposable point-of-need
device will improve the speed and efficacy of triage and treatment and enhance the
standard of

Full Details: The DoD requires timely and comprehensive threat detection to sup-
port overall readiness, counter the spread of disease, and promote stabilization mis-
sions. State of the art diagnostic and biosurveillance systems are unable to keep pace
with disease outbreaks and fail to support decision-making at the time and place of
need. The “Detect It with Gene Editing Technologies” (DIGET) program aims to
leverage advances in gene editing technologies to develop field-forward diagnostic and
biosurveillance technologies that enable detection of any threat, anytime, anywhere.
To achieve its goal, DIGET performers will design, develop, prototype, and deploy
two novel nucleic acid detection devices: 1) a disposable point-of-need diagnostic for
up to 10 pathogens and host biomarkers, and 2) a massively multiplexed detection
(MMD) device for 1,000 or more targets. Both devices must be simple to operate,
low-cost, and rapidly reconfigurable to provide high-impact, high-quality, and trusted
information that enhances decision-making. If successful, the disposable point-of-need
device will improve the speed and efficacy of triage and treatment and enhance the
standard of care for the military and public health domains in austere environments.
The MMD device will enable early threat detection, assess disease severity, and im-
prove situational awareness. The MMD platform will also provide actionable data for
biosurveillance efforts such as characterization of known and emergent pathogens in
circulation to inform countermeasure deployment.

228
Data Protection in Virtual Environ-
ments (DPRIVE)
Dr. Bryan Jacobs

https://www.darpa.mil/program/data-protection-in-virtual-environments.html

GPT-3 Summary: GOAL: The DPRIVE program will enable fast and secure com-
putation on encrypted data.
TECHNOLOGY METHODS: Design and develop a hardware accelerator that will
dramatically reduce the processing overhead required to implement FHE calculations.
IMPACT: Reduce the computational run time overhead by many orders of magnitude
compared to current software-based FHE computations on conventional CPUs.

Full Details: The safety and security of critical information – whether it is sensitive
intellectual property (IP), financial information, personally identifiable information
(PII), intelligence insight, or beyond – is of vital importance. Current methods pro-
tect data as it is transmitted across a network, at rest, or while in storage. Processing
or computing on this data, however, requires that it is first decrypted, exposing it to
numerous vulnerabilities and threats. Fully homomorphic encryption (FHE) offers a
solution to this challenge. FHE enables computation on encrypted data, or ciphertext,
to keep data protected at all times. The benefits of FHE are significant, from enabling
the use of untrusted networks to enhancing data privacy. Despite its potential, FHE
requires enormous computation time to perform even simple operations, making it ex-
ceedingly impractical to implement with traditional processing hardware. The Data
Protection in Virtual Environments (DPRIVE) program seeks to enable FHE compu-
tation within a factor of ten of unencrypted computations, enabling data security for
all states of data across DoD and commercial applications.While a number of FHE
solutions have been developed, running FHE in software on standard processing hard-
ware remains untenable for practical data security applications due to the massive
processing overhead. DPRIVE aims to design and develop a hardware accelerator
integrating novel architectural approaches, hardware, and software to dramatically
reduce the processing overhead required to implement FHE calculations. Specifically,
the program seeks to develop novel approaches to memory management, flexible data
structures and programming models, and formal verification methods that ensure the
design of the FHE implementation is effective and accurate, while also dramatically
decreasing the overhead penalty required for FHE computations. The target accelera-
tor should reduce the computational run time overhead by many orders of magnitude
compared to current software-based FHE computations on conventional CPUs, and

229
accelerate FHE calculations to within one order of magnitude of current performance
on unencrypted data.

230
Phorcys
Mr. John Waterston

https://www.darpa.mil/program/phorcys.html

GPT-3 Summary: The goal of the Phorcys program is to develop a new form
of access for military operations through the coordination of containerized military
capabilities distributed across available commercial vessels. The Phorcys software is
agnostic of payload and optimizes the execution of a variety of mission sets across
wide regions of interest. This could potentially reduce operational costs and increase
the speed of deployment for the military.

Full Details: Phorcys seeks to enable a new form of access for military operations
through coordination of containerized military capabilities distributed across available
commercial vessels. Utilizing the capacity of commercial vessels alleviates the current
need for “exquisite” military vessels, reduces operational costs, and is more rapidly
deployable. Phorcys’ mosaic maritime planning and orchestration software is agnostic
of payload and optimizes the execution of a variety of mission sets across wide regions
of interest.

231
Common Heterogeneous Integration
and IP Reuse Strategies (CHIPS)
Dr. Gordon Keeler

https://www.darpa.mil/program/common-heterogeneous-integration-and-ip-reuse-strategies.html

GPT-3 Summary: GOAL: The CHIPS program seeks to establish a new paradigm
in IP reuse by creating an ecosystem of discrete modular, reusable IP blocks.
TECHNOLOGY METHODS: Develop the design tools and integration standards
required to demonstrate modular IC designs.
IMPACT: Create a system of ICs that leverages the best of DoD and commercial
designs and technologies.

Full Details: The explosive growth in mobile and telecommunication markets has
pushed the semiconductor industry toward integration of digital, analog, and mixed-
signal blocks into system-on-chip (SoC) solutions. Advanced silicon (Si) complemen-
tary metal oxide semiconductor (CMOS) technology has enabled this integration,
but has also led to a rise in costs associated with design and processing. Driven
by aggressive digital CMOS scaling for high-volume products, Intellectual Property
(IP) reuse has emerged as a tool to help lower design costs associated with advanced
SoCs. The monolithic nature of state-of-the-art SoCs is not always acceptable for
Department of Defense (DoD) or other low-volume applications due to factors such
as high initial prototype costs and requirements for alternative material sets. To en-
hance overall system flexibility and reduce design time for next-generation products,
the Common Heterogeneous Integration and Intellectual Property (IP) Reuse Strate-
gies (CHIPS) program seeks to establish a new paradigm in IP reuse. The vision of
CHIPS is an ecosystem of discrete modular, reusable IP blocks, which can be assem-
bled into a system using existing and emerging integration technologies. Modularity
and reusability of IP blocks will require electrical and physical interface standards to
be widely adopted by the community supporting the CHIPS ecosystem. Therefore,
the CHIPS program will develop the design tools and integration standards required
to demonstrate modular integrated circuit (IC) designs that leverage the best of DoD
and commercial designs and technologies.

232
Symbiotic Design for Cyber Physi-
cal Systems
Dr. Matthew Wilding

https://www.darpa.mil/program/symbiotic-design-for-cyber-physical-systems.html

GPT-3 Summary: GOAL: The Symbiotic Design for Cyber Physical Systems
(SDCPS) program will develop AI-based approaches to enable the correct-by-
construction design of military-relevant CPS.

METHODS: Address three intrinsic challenges of CPS design: predictability,

convergence, and exploration.

IMPACT: Reduce the time from CPS inception to deployment from years to
months, and enhance innovation in design.

Full Details: Cyber physical systems (CPS) are instrumental to current and future
Department of Defense (DoD) mission needs – unmanned vehicles, weapon systems,
and mission platforms are all examples of military-relevant CPS. These systems and
platforms integrate cyber and physical subsystems, and the enormous complexity of
the resulting CPS has made their engineering design a daunting challenge. An imme-
diate consequence of this complexity is development cycles with prolonged timelines
that challenge DoD’s ability to counter emerging threats.CPS design is a complex en-
deavor that involves many domains – from cyber (e.g., software, control, computing,
and communication) to physical (e.g., structural, mechanical, thermal, etc.) to man-
ufacturing – and upwards of hundreds of domain-specific tools orchestrated by large
teams of engineers with extensive domain knowledge and subject matter expertise.
Current engineering design processes start with requirements-driven decomposition
into discipline-specific design flows that, at their core, are concurrently running se-
quential decision-making processes that involve generating candidate architectures;
evaluating, selecting, and refining options; and integrating the design until require-
ments are satisfied. The goal of the Symbiotic Design for CPS (SDCPS) program is
to develop AI-based approaches to enable correct-by-construction design of military-
relevant CPS. SDCPS seeks to reduce the time from CPS inception to deployment
from years to months, and enhance innovation in design. To accomplish this, SDCPS
will address the following three intrinsic challenges: Predictability – The soundness of
design decisions relies on accurate predictions of performance prior to the implemen-
tation of software and physical components. However, accurate predictions require

233
high-fidelity models that are cost- and time-prohibitive to produce. Cost-effective
modeling processes produce results with substantial uncertainty.Convergence – De-
sign teams are federated according to discipline boundaries. However, separation of
concerns within a complex system neglects the interdependence of design decisions,
rendering rapid convergence to a viable integrated solution practically impossible.
Exploration – Limited by time and resources, engineers have to make tradeoffs that
constrain the exploration of the design space to the familiar and known-feasible. This
leaves vast areas of design space unexplored, which may contain unconventional but
highly performant solutions.The vision of the program is to vastly expand coverage
and accelerate exploration of CPS design spaces with the symbiosis of two very differ-
ent kinds of agents: humans with their uncanny ability to create intuitive associations
across design domains, and machines with their ability to recognize statistical pat-
terns from data and navigate vast search spaces for optimal solutions. The program
aims to realize this vision by transforming the human-focused model-based design
flows used today into a symbiotic process of collaborative discovery by humans and
continuously learning AI-based co-designers. For more information, please see the
Symbiotic Design for Cyber Physical Systems Proposers Day Announcement and the
Symbiotic Design for Cyber Physical Systems BAA.

234
AdvaNced airCraft Infrastructure-
Less Launch And RecoverY (AN-
CILLARY)
Mr. Steve Komadina

https://www.darpa.mil/program/advanced-aircraft-infrastructure-less-launch-and-recovery.html

GPT-3 Summary: GOAL: The ANCILLARY program will develop and demon-
strate an X-plane with the critical technologies required for a leap-ahead in long
endurance, vertical takeoff and landing (VTOL) unmanned air system (UAS) perfor-
mance.
TECHNOLOGY METHODS: The UAS would be able to launch and recover from
ship flight decks and small austere land locations in adverse weather without addi-
tional infrastructure equipment, thus enabling expeditionary deployments.
IMPACT: The small UAS size would allow many aircraft to be stored and operated
from one ship creating a tactical beyond-line-of-site (BL

Full Details: The AdvaNced airCraft Infrastructure-Less Launch And RecoverY

(ANCILLARY) program aims to develop and flight demonstrate an X-plane with the
critical technologies required for a leap-ahead in long endurance, vertical takeoff and
landing (VTOL) unmanned air system (UAS) performance. The UAS would be able
to launch and recover from ship flight decks and small austere land locations in adverse
weather without additional infrastructure equipment, thus enabling expeditionary
deployments. Unlike large VTOL systems, the small UAS size would allow many
aircraft to be stored and operated from one ship creating a tactical beyond-line-of-
site (BLOS) multi-intelligence sensor network capability.

235
Cyber Assured Systems Engineer-
ing (CASE)
Mr. William Martin

https://www.darpa.mil/program/cyber-assured-systems-engineering.html

GPT-3 Summary: GOAL: The goal of CASE is to develop the necessary design,
analysis and verification tools to allow system engineers to design-in cyber resiliency
and manage tradeoffs as they do other nonfunctional properties when designing
complex embedded computing systems.

TECHNOLOGY METHODS: The program will research breakthroughs in:

the elicitation of cyber resiliency requirements before the system is built; the design
and verification of systems when requirements are not testable (i.e., when they are
expressed in shall not statements); tools to automatically adapt software to new
non-functional requirements; and techniques to scale

Full Details: Embedded computing systems are ubiquitous in critical infrastruc-

ture, vehicles, smart devices, and military systems. Conventional wisdom once held
that cyberattacks against embedded systems were not a concern since they seldom
had traditional networking connections on which an attack could occur. However,
attackers have learned to bridge air gaps that surround the most sensitive embedded
systems, and network connectivity is now being extended to even the most remote
of embedded systems. In short, embedded systems are now subject to cyberattacks,
either as the end goal of the cyber assailant or as a means to a greater end, and there
is a critical need to protect and defend embedded systems in a cyber-context. The
mechanisms currently employed to secure embedded systems include development
of software using cyber best practices, adapting mechanisms from information tech-
nology (IT) systems, and penetration testing followed by patching. Unfortunately,
these methods have proven to be generally ineffective. Critical systems are built by
requirements-based engineering and it is an accepted axiom of systems engineering
that requirements are positive, testable statements about the system—statements on
the systems’ functional behaviors and non-functional properties often captured as
“shall” statements. This style of engineering has proven to be ineffective in engineer-
ing cyber resilient systems because cyber requirements are often statements on what
the system should not do, i.e., “shall not” statements. The goal of CASE is to de-
velop the necessary design, analysis and verification tools to allow system engineers
to design-in cyber resiliency and manage tradeoffs as they do other nonfunctional

236
properties when designing complex embedded computing systems. Cyber resiliency
means the system is tolerant to cyberattacks in the same way that safety critical
systems are tolerant to random faults—they recover and continue to execute their
mission function. Achieving this goal requires research breakthroughs in:the elicita-
tion of cyber resiliency requirements before the system is built; the design and veri-
fication of systems when requirements are not testable (i.e., when they are expressed
in shall not statements);tools to automatically adapt software to new non-functional
requirements; andtechniques to scale and provide meaningful feedback from analysis
tools that reside low in the development tool chain.The CASE BAA is available at
https://www.fbo.gov/index?s=opportunityandmode=formandid=3353e66b84ebc0e87b2e5e263c3da5a
cview=1

237
RNET
Dr. Richard Singer

https://www.darpa.mil/program/rnet.html

GPT-3 Summary: GOAL: The RNET program seeks to demonstrate RF com-

munications and sensing technologies that enable performance equivalent to existing
systems, but on an order-of-magnitude smaller platform.
TECHNOLOGY METHODS: The program is interested in developing lightweight,
deployable antennas and advanced software defined radios.
IMPACT: If successful, the program could enable the development of smaller, more
agile satellites that are less expensive to launch and maintain.

Full Details: Due to advancements in component technology, microsatellite sys-

tems are increasingly viable solutions to address earth science and remote sensing
missions. For example, constellations of commercial, small, optical satellites are pro-
liferating, supporting a variety of tasks and data applications. The ability to eco-
nomically launch microsatellites on diverse launch vehicles promises advantages in
rapid technology refresh, responsive space operations, and resilient (i.e., redundant)
systems. The RNET program seeks to demonstrate RF communications and sensing
technologies that enable performance equivalent to existing systems, but on an order-
of-magnitude smaller platform. Two key technology areas of interest are lightweight,
deployable antennas and advanced software defined radios.

238
Tailorable Feedstock and Forming (TFF)
Dr. William Carter

https://www.darpa.mil/program/tailorable-feedstock-and-forming.html

GPT-3 Summary: GOAL: Develop a new composite material format (feedstock)

and processing technologies to reduce manufacturing complexity and enable use of
advanced materials for small parts at a competitive cost.

TECHNOLOGY METHODS: Develop a new composite material feedstock and as-

sociated processing technologies to enable the manufacturing of advanced small parts.

IMPACT: This new program would enable greater efficiency in manufacturing

advanced small parts with decreased development and production costs.

Full Details: The capabilities and technical specifications required for Department of
Defense (DoD) platforms are constantly changing due to unanticipated circumstances,
needs and emerging threats. However, complex development and design cycles and the
associated high costs of structural design changes for current technologies significantly
limit our ability to rapidly and affordably evolve such systems. The DARPA TFF
program will address these challenges by seeking to develop a new composite material
format (i.e., feedstock) and associated processing technologies (e.g., reconfigurable
forming) to reduce manufacturing complexity and enable use of advanced materials
for small parts weighing less than 20 pounds at costs competitive with aluminum.
These technologies would enable greater efficiency in manufacturing advanced DoD
small composite parts with decreased development and production costs.

239
Automated Rapid Certification Of
Software (ARCOS)
Mr. William Martin

https://www.darpa.mil/program/automated-rapid-certification-of-software.html

GPT-3 Summary: GOAL: Automate the evaluation of software assurance evidence

to speed up the certification process.
TECHNOLOGY METHODS: Use model-based design technology and probabilistic
model checking to reason over design and quantify uncertainty. Use ”Big Code”
analytics to understand software and its associated artifacts. Use assurance case
languages to express arguments on how software meets certification goals.
IMPACT: Improve the efficiency of the software certification process by automating
evidence generation, curation, and assessment.

Full Details: The process of determining that a software system’s risk is acceptable
is referred to as “certification.” Current certification practices within the Department
of Defense (DoD) are antiquated and unable to scale with the amount of software
deployed. Two factors prevent scaling: (a) the use of human evaluators to determine
if the system meets certification criteria, and (b) the lack of a principled means to de-
compose evaluations.The amount of assurance evidence needed to determine software
conformance to certification can be overwhelming to human subject matter experts,
resulting in superficial, incomplete, and/or unacceptably long evaluations. Human
evaluators also have unique expertise, experience, and biases that influence their ap-
proach to evaluations. Because certification requirements may be vague or poorly
written, evaluators often must interpret what is intended. Combined, these factors
result in inconsistencies over time and across evaluations. Currently, there is no means
to compose evaluations in a principled and trustworthy manner. Composed evalu-
ations would allow subsystems or components to be evaluated independently. The
results of those independent evaluations could then be leveraged as assurance evidence
in the composed systems. This would amortize the effort of evaluating any compo-
nent over all systems using that component. Current practice requires re-evaluating
components and their assurance evidence in every system that employs them. The in-
ability to use a divide–and-conquer approach to certification of large systems increases
the cost and time required to perform these certifications.The goal of the Automated
Rapid Certification Of Software (ARCOS) program is to automate the evaluation of
software assurance evidence to enable certifiers to determine rapidly that system risk
is acceptable. Two factors support the acceleration of software certification through

240
the automation of evaluations. First, the DoD has articulated its intentions to have
its contractors modernize their engineering processes in the DoD Digital Engineering
Strategy. The goal of this strategy is to move away from document-based engineering
processes and towards design models that are to be the authoritative source of truth
for systems. Such a future does not lend itself to current certification practices, but
it will facilitate the automated evaluation of assurance. Second, advances in several
technologies provide a basis for confidence that automated evaluation of assurance
evidence to support certification is possible. Model-based design technology, includ-
ing probabilistic model checking, may enable reasoning over a design in a way that
quantifies uncertainty. So-called “Big Code” analytics have pioneered the application
of semantic-based analytics to software and its associated artifacts. Mathematically
rigorous analysis and verification provide the ability to develop software implementa-
tions that are demonstrably correct and sound. Assurance case languages provide us
a means for expressing arguments on how software fulfills its certification goals, in a
manner that is machine-readable.ARCOS will explore techniques for automating the
evidence generation process for new and legacy software; create a means of curating
evidence while maintaining its provenance; and develop technologies for the auto-
mated construction of assurance cases, as well as technologies that can validate and
assess the confidence of an assurance case argument. The evidence generation, cura-
tion, and assessment technologies will form the ARCOS tools and processes, working
collectively to provide a scalable means of accelerating the pathway to certification.For
more information please visit the Automated Rapid Certification of Software (AR-
COS) Proposers Day webpage available here. For a video of the ARCOS Proposers
Day presentations, click here.

241
Investigating Adaptive Modal Bases
for Intelligent Classification (IAMBIC)
Dr. Joe Altepeter

https://www.darpa.mil/program/investigating-adaptive-modal-bases-for-intelligent-classification.htm

GPT-3 Summary: GOAL: The goal of this program is to develop a theoretical

framework for adaptively varying measurement bases in response to initial measure-
ment data.
METHODS: The performers will develop a cohesive and general theoretical frame-
work for adaptively varying measurement bases in response to initial measurement
data. They will apply their adaptive measurement framework on two (or more) tasks:
The first task will address the DARPA-defined Constellation Challenge problem of
imaging groups of unresolved sources beyond the Rayleigh resolution limit, in order
to quantitatively bound the conditions under which adaptive modal imaging provides
a positive real- world impact for the specific problem of sub-

Full Details: Recent imaging techniques inspired by the mathematics of quan-

tum information claim to enable imaging beyond the traditional Rayleigh resolution
limit.The approach is an example of “modal imaging” and is just one recent example
of a much broader group of techniques developed over many decades that attempt to
efficiently extract information from multidimensional systems using non-Fourier mea-
surement bases. In almost all of these cases, prior information enables more efficient
or more accurate information recovery from measurement results. More importantly,
even without prior information, intelligently and continually adapting the measure-
ment basis in response to earlier measurement results enables superior performance for
many information retrieval tasks.The Investigating Adaptive Modal Bases for Intelli-
gent Classification (IAMBIC) Disruption Opportunity (DO) is an effort to create new
frameworks for analyzing when such adaptive measurement bases can provide signifi-
cant, real-world advantage for important problems.IAMBIC performers will develop a
cohesive and general theoretical framework for adaptively varying measurement bases
in response to initial measurement data. They will apply their adaptive measurement
framework on two (or more) tasks: The first task will address the DARPA-defined
Constellation Challenge problem of imaging groups of unresolved sources beyond
the Rayleigh resolution limit, in order to quantitatively bound the conditions under
which adaptive modal imaging provides a positive real- world impact for the spe-
cific problem of sub-Rayleigh-resolution imaging. The other measurement tasks are
performer-defined. The performers will analyze these tasks using their adaptive mea-

242
surement framework and show that their strategy provides a significant real-world
impact.

243
OFFensive Swarm-Enabled Tactics
(OFFSET)
Dr. Lael Rudd

https://www.darpa.mil/program/offensive-swarm-enabled-tactics.html

GPT-3 Summary: The goal of the OFFSET program is to develop swarm technolo-
gies to enable small-unit infantry forces to accomplish diverse missions in complex ur-
ban environments. The technology methods include developing an advanced human-
swarm interface, a real-time, networked virtual environment, and a community-driven
swarm tactics exchange. The impact of the program is to provide the tools to quickly
generate swarm tactics, evaluate those swarm tactics for effectiveness, and integrate
the best swarm tactics into field operations.

Full Details: DARPA’s OFFensive Swarm-Enabled Tactics (OFFSET) program

envisions future small-unit infantry forces using swarms comprising upwards of 250
small unmanned aircraft systems (UASs) and/or small unmanned ground systems
(UGSs) to accomplish diverse missions in complex urban environments. By lever-
aging and combining emerging technologies in swarm autonomy and human-swarm
teaming, the program seeks to enable rapid development and deployment of break-
through capabilities. OFFSET aims to provide the tools to quickly generate swarm
tactics, evaluate those swarm tactics for effectiveness, and integrate the best swarm
tactics into field operations. To accomplish these goals, OFFSET will develop an
active swarm tactics development ecosystem and supporting open systems architec-
ture, including: An advanced human-swarm interface to enable users to monitor
and direct potentially hundreds of unmanned platforms simultaneously in real time.
The program intends to leverage rapidly emerging immersive and intuitive interactive
technologies (e.g., augmented and virtual reality, voice-, gesture-, and touch-based) to
create a novel command interface with immersive situational awareness and decision
presentation capabilities. The interface would also incorporate a swarm interaction
grammar, enabling “freestyle” design of swarm tactics that allow dynamic action and
reaction based on real-time conditions in the field. A real-time, networked virtual
environment that would support a physics-based, swarm tactics game. In the game,
players would use the interface to rapidly explore, evolve, and evaluate swarm tactics
to see which would potentially work best, whether using various unmanned systems
in the real world or exploring innovative synthetic technologies in the virtual one.
Users could submit swarm tactics and track their performance from test rounds on a
leaderboard, to encourage both competition and collaboration. A community-driven

244
swarm tactics exchange. This curated, limited access program portal would house
tools to help participants design swarm tactics by composing collective behaviors,
swarm algorithms, and even other existing tactics. It would provide these key in-
gredients to an extensible architecture for end-user-generated swarm tactics and help
create a lasting community to innovate and cultivate the most effective tactics, with
the potential to integrate third-party tactics and future users. OFFSET is structured
to demonstrate its technologies through frequent live experiments with various un-
manned air and ground platforms. Roughly every six months, operational vignettes
of progressively increasing complexity would challenge both the swarm architecture
and the developed swarm tactics across numerous technological and operational test
variables, such as swarm size, spatial scale of operations, and mission duration. Users
would employ the swarm interface to test the best of the virtual tactics in the real
world, and interactively supply their unmanned platforms with near-real-time tactics
updates using automated deployment technologies.

245
Resilient Supply-and-Demand Net-
works (RSDN)
Dr. Mark Flood

https://www.darpa.mil/program/resilient-supply-and-demand-networks.html

GPT-3 Summary: GOAL: Improve the resilience of the Department of Defense’s

(DoD) supply-and-demand networks (SDNs) in the face of disruptions.
TECHNOLOGY METHODS: Develop novel information resources and practical an-
alytical tools, including granular maps of SDNs, to expose and mitigate sources of
surprise.
IMPACT: Improve the DoD’s ability to respond to disruptions in its supply-and-
demand networks.

Full Details: The Department of Defense (DoD) has a critical need to secure its
sources of materiel against both intentional—including adversarial—and uninten-
tional disruptions. Extensive global networks of private-sector vendors, commonly
called “supply chains,” collaborate to provide these key resources, including precur-
sor components and materials. The Resilient Supply-and-Demand Networks (RSDN)
program adopts the phrase “supply-and-demand network” (SDN) in lieu of “supply
chain” to emphasize that the strategic challenges are more extensive than the logistic
challenges of delivering (“supplying”) materiel. SDNs are open, complex, evolving
systems whose dynamics reflect the impact of both external factors (e.g., conflict,
climate change) and internal behaviors (e.g., inventory management). These factors
are often driven by the locally focused decisions of SDN participants themselves. The
economic forces of globalization have historically emphasized SDN efficiency (cost and
timeliness) over SDN resilience (response to shocks). SDN resilience is a characteristic
of the SDN system as a whole. Resilience enhancements, therefore, require coordi-
nated action among SDN participants, who may individually lack the incentive and
discretion to undertake this complex process. SDNs are potential sources of strategic
surprise for the DoD. Surprise in the context of SDNs is rarely beneficial. Key under-
lying sources of SDN surprise are:Imperfect knowledge of the current structure of the
SDN itself;Difficulty in predicting different kinds of shocks from the wide landscape
of threats and vulnerabilities; and, Difficulty in predicting changes in the SDN due
to endogenous behavioral shifts, which can be amplified abruptly and dramatically
through feedback effects.To expose and mitigate these sources of surprise, RSDN will
develop novel information resources and practical analytical tools, including:Granular,
up-to-date maps of SDNs, identifying individual providers (network nodes) and their

246
specific trading relationships with each other (network edges);Augmentation of the
basic network maps with detailed organizational features describing the nodes and
detailed descriptions of the procurement relationships governing the edges;A broad
and extensible set of analytical tools for exploring fragilities (threats and vulnerabil-
ities) in SDNs; and,Techniques for stress testing SDNs by specifying and simulating
their responses to exogenous shocks and endogenous feedback.Additional information
is available in the RSDN BAA.

247
ReVector
Dr. Linda Chrisey

https://www.darpa.mil/program/revector.html

GPT-3 Summary: GOAL: To reduce attraction and feeding by mosquitoes by

modulating the profile of skin-associated volatile molecules.
METHODS: Changing the organisms that are present in the skin microbiome and/or
their metabolic processes.
IMPACT: Improve protection against disease vectors for military personnel.

Full Details: Mosquitoes transmit pathogens that cause dengue, malaria, and other
diseases that present significant risks to the readiness and resilience of military per-
sonnel, and public health more generally. The ReVector program aims to maintain
the health of military personnel operating in disease-endemic regions by reducing at-
traction and feeding by mosquitoes. Mosquitoes are attracted to the general area of
humans by volatiles emitted in human breath. However, it is the heat and volatile
molecules from human skin that direct mosquitoes to the specific sites on the body
where they feed. Many of those volatile molecules are produced by the metabolism of
organisms in the skin microbiome. Researchers on the ReVector program are working
to develop precise, safe, and efficacious technologies to modulate the profile of skin-
associated volatile molecules by changing the organisms that are present in the skin
microbiome and/or their metabolic processes. Although other approaches already
exist to slow the spread of vector-borne disease (e.g., bed nets, chemical repellants,
anti-malarial drugs), they each have logistical burdens or side effects that make them
impractical for use during military deployments. For example, the requirements for
frequent reapplication of repellants or repeat dosing of drugs often result in incon-
sistent protection. In contrast, the envisioned ReVector treatments could be applied
just hours before a mission with minimal equipment or training, would produce no
detectable odor, and would last for up to two weeks without reapplication, offering
improved, sustained protection against disease vectors.

248
Warfighter Analytics using Smart-
phones for Health (WASH)
Mr. Tejas Patel

https://www.darpa.mil/program/warfighter-analytics-using-smartphones-for-health.html

GPT-3 Summary: The goal of the WASH program is to use data collected from
cellphone sensors to enable algorithms that can conduct passive, continuous, real-time
assessment of the warfighter. The objective is to extract physiological signals from
the data that may help determine the warfighter’s current health status and identify
latent or developing health disorders. The program is expected to research the devel-
opment of algorithms and techniques for identifying known indicators of physiological
problems and deviations from the warfighter’s micro-behaviors that could indicate
such problems. Additionally, the program may identify new ”digital biomarkers” of
physiological problems through the combination of big data analytics and

Full Details: Currently, understanding and assessing the readiness of the warfighter
involves medical intervention with the help of advanced equipment, such as electrocar-
diographs (EKGs) and other specialized medical devices, that are too expensive and
cumbersome to employ continuously or without supervision in non-controlled environ-
ments. On the other hand, currently 92 percent of adults in the United States own a
cell phone, which could be used as the basis for continuous, passive health, and readi-
ness assessment. The Warfighter Analytics using Smartphones for Health (WASH)
program seeks to use data collected from cellphone sensors to enable novel algorithms
that conduct passive, continuous, real-time assessment of the warfighter. The ob-
jective of WASH is to extract physiological signals, which may be weak and noisy,
that are embedded in the data obtained through existing mobile device sensors (e.g.,
accelerometer, screen, microphone). Such extraction and analysis, done on a continu-
ous basis, may help determine current health status and identify latent or developing
health disorders (Figure 1). Figure 1: WASH health determination WASH research
will explore the development of algorithms and techniques for identifying both known
indicators of physiological problems (such as disease, illness, and/or injury) and devi-
ations from the warfighter’s micro-behaviors that could indicate such problems. It is
also expected that additional “digital biomarkers” of physiological problems may be
identified during the research through the combination of big data analytics and med-
ical ground truth provided to performers. Digital biomarkers are consumer-generated
physiological and behavioral measures collected through connected digital tools, in
this case a smartphone. A prerequisite for the extraction and interpretation of the raw

249
sensor data and any identified digital biomarkers is determining the context of such
data collection and analysis, which may affect the relevance of any given sensor and
permit “denoising,” or elimination of irrelevant or misleading readings. For example,
relying on cellphone accelerometer data while the warfighter is in a moving vehicle
would likely negatively influence the utility of such data for WASH-type analysis un-
less the auxiliary motion is identified and cancelled. Thus, key focus areas will be
the extraction of the signal context and the identification of complicated actions and
environmental variables, and the association of user state with symptoms of illness
conditions in order to identify potential illnesses and conditions before conventional
symptomatic display. It is the union of personal behavior/characteristics, smartphone
sensor collection, context of smartphone use, and disease biomarkers that will define
the preclinical health determination of the WASH program (Figure 2). Figure 2:
WASH program concept The program goal is to enable the creation of a mobile ap-
plication that passively assesses a warfighter’s readiness immediately and over time.
This application seeks to provide: Clinicians with plausible health conditions sup-
ported by the analysis, determination, and fusion of digital biomarkers for disease
correlation against ground truth; Commanders with unit readiness information, both
at the current time and in the near future; and Users of the device with information
about their current health status and early indicators of medical conditions.

250
Guaranteeing AI Robustness Against
Deception (GARD)
Dr. Bruce Draper

https://www.darpa.mil/program/guaranteeing-ai-robustness-against-deception.html

GPT-3 Summary: GOAL: The goal of the GARD program is to create theoretical
foundations to identify vulnerabilities in machine learning systems, and to create
effective defenses against a broad range of attacks.

METHODS: The GARD program will use a novel testbed to evaluate AI de-
fenses in a variety of scenarios.

IMPACT: The goal of the program is to create more robust machine learning
systems that are less vulnerable to attack.

Full Details: The growing sophistication and ubiquity of machine learning (ML)
components in advanced systems dramatically expands capabilities, but also increases
the potential for new vulnerabilities. Current research on adversarial AI focuses on
approaches where imperceptible perturbations to ML inputs could deceive an ML
classifier, altering its response. Such results have initiated a rapidly proliferating field
of research characterized by ever more complex attacks that require progressively less
knowledge about the ML system being attacked, while proving increasingly strong
against defensive countermeasures. Although the field of adversarial AI is relatively
young, dozens of attacks and defenses have already been proposed, and at present
a comprehensive theoretical understanding of ML vulnerabilities is lacking. GARD
seeks to establish theoretical ML system foundations to identify system vulnerabili-
ties, characterize properties that will enhance system robustness, and encourage the
creation of effective defenses. Currently, ML defenses tend to be highly specific and
are effective only against particular attacks. GARD seeks to develop defenses capa-
ble of defending against broad categories of attacks. Furthermore, current evaluation
paradigms of AI robustness often focus on simplistic measures that may not be rele-
vant to security. To verify relevance to security and wide applicability, defenses gener-
ated under GARD will be measured in a novel testbed employing scenario-based eval-
uations. As part of the program, GARD researchers from Two Six Technologies, IBM,
MITRE, University of Chicago, and Google Research generated the following virtual
testbed, toolbox, benchmarking dataset, and training materials that are now available
to broader research community:The Armory virtual platform, available on GitHub,

251
serves as a “testbed” for researchers in need of repeatable, scalable, and robust eval-
uations of adversarial defenses.Adversarial Robustness Toolbox (ART) provides tools
for developers and researchers to defend and evaluate their ML models and applica-
tions against a number of adversarial threats.The Adversarial Patches Rearranged In
COnText (APRICOT) dataset enables reproducible research on the real-world effec-
tiveness of physical adversarial patch attacks on object detection systems.The Google
Research Self-Study repository contains “test dummies” that represent a common
idea or approach to build defenses.The GARD program’s Holistic Evaluation of Ad-
versarial Defenses repository is available at https://www.gardproject.org/. Interested
researchers are encouraged to take advantage of these resources and check back often
for updates.

252
Science of Atomic Vapors for New
Technologies (SAVaNT)
Dr. Tatjana Curcic

https://www.darpa.mil/program/science-of-atomic-vapors-for-new-technologies.html

GPT-3 Summary: GOAL: The SAVaNT program seeks to develop atomic vapors
as a platform for new technologies in electric field sensing and imaging, magnetic field
sensing, and quantum information science.
TECHNOLOGY METHODS: The program will focus on improving atomic coherence
at room temperature in order to advance Rydberg electrometry, vector magnetometry,
and vapor quantum electrodynamics.
IMPACT: If successful, the SAVaNT program could lead to new technologies with a
variety of applications, including field measurements with low size, weight, and power
requirements, as well as scalable room

Full Details: The “Science of Atomic Vapors for New Technologies” (SAVaNT)
program aims to significantly advance the performance of atomic vapors as a room-
temperature (RT) platform for enabling new technologies in the areas of electric field
sensing and imaging, magnetic field sensing, and quantum information science (QIS).
SAVaNT has three Technical Areas based on the application domains where atomic
vapors are expected to have the biggest impact: (1) Rydberg Electrometry, (2) Vector
Magnetometry, and (3) Vapor Quantum Electrodynamics (vQED). The common sci-
entific challenge is to improve atomic coherence at room temperature. The associated
technical challenges depend somewhat on the specifics of the most dominant deco-
herence mechanisms in each approach. Program goals include improving sensitivity
of Rydberg electrometry, developing vector modality for atomic magnetometry with
high sensitivity and accuracy, and demonstrating cavity quantum electrodynamics
with atomic vapors. If successful, SAVaNT advances will lay the foundations for new
technologies that address important DoD needs, including applications that require
low size, weight, and power (SWaP), high sensitivity electric and magnetic field mea-
surements as well as applications that require scalable room-temperature quantum
memories and interfaces.

253
Assured Neuro Symbolic Learning
and Reasoning (ANSR)
Dr. Alvaro Velasquez

https://www.darpa.mil/program/assured-neuro-symbolic-learning-and-reasoning.html

GPT-3 Summary: GOAL: The goal of the ANSR program is to create trustworthy
autonomous systems through hybrid AI algorithms.
TECHNOLOGY METHODS: The program will integrate symbolic reasoning with
data-driven learning to create robust, assured, and therefore trustworthy systems.
IMPACT: The impact of this program will be the development of autonomous systems
that are safe and reliable.

Full Details: To successfully incorporate autonomous systems into their missions,

military operators must have confidence that those systems will operate safely and
perform as intended.DARPA is motivating new thinking and approaches to artificial
intelligence development to enable high levels of trust in autonomous systems through
the Assured Neuro Symbolic Learning and Reasoning (ANSR) program.ANSR seeks
breakthrough innovations in the form of new, hybrid AI algorithms that integrate
symbolic reasoning with data-driven learning to create robust, assured, and there-
fore trustworthy systems. ANSR defines a system as trustworthy, if it is: Ro-
bust to domain-informed and adversarial perturbations;Supported by an assurance
framework that creates and analyzes heterogenous evidence towards safety and risk
assessments; and Predictable with respect to some specification and models of fit-
ness.Advances in assurance technologies, including formal and simulation-based ap-
proaches, have helped accelerate identification of failure modes and defects of ma-
chine learning (ML) algorithms. Unfortunately, the ability to repair defects in state-
of-the-art ML remains limited to retraining, which is not guaranteed to eliminate
defects or to improve the generalizability of ML algorithms. Further, while the
runtime assurance architecture (e.g. monitoring and recovery) ensures operational
safety, frequent invocations of fallback recovery, triggered by brittleness and gen-
eralizability of ML, compromises the ability to accomplish a mission. ANSR hy-
pothesizes that several of the limitations in ML today are a consequence of the
inability to incorporate contextual and background knowledge, and treating each
data set as an independent, uncorrelated input. In the real world, observations are
often correlated and a product of an underlying causal mechanism, which can be
modeled and understood. ANSR also posits that hybrid AI algorithms capable of
acquiring and integrating symbolic knowledge and performing symbolic reasoning

254
at scale will deliver robust inference, generalize to new situations, and provide evi-
dence for assurance and trust.The ANSR Broad Agency Announcement is available
at https://sam.gov/opp/59cc5fec16c24ae0a2838b1b21e263a2/view

255
Ice Control for cold Environments
(ICE)
Dr. Anne Cheever

https://www.darpa.mil/program/ice-control-for-cold-environments.html

GPT-3 Summary: The ICE program seeks to develop solutions for extreme
cold weather DOD operational challenges by discovering and optimizing biologically
sourced or inspired molecules to enable tuned inhibition or nucleation of ice crystal-
lization, propagation, and adhesion. The best performing molecules identified during
the discovery phase will subsequently be formulated to develop novel materials with
enhanced operational performance, tailored to mitigate extreme cold weather chal-
lenges for specific DOD applications, and tested based on stakeholder needs and
priorities.

Full Details: The Ice Control for cold Environments (ICE) program seeks to develop
solutions for extreme cold weather DOD operational challenges by discovering and
optimizing biologically sourced or inspired molecules to enable tuned inhibition or
nucleation of ice crystallization, propagation, and adhesion. The best performing
molecules identified during the discovery phase will subsequently be formulated to
develop novel materials with enhanced operational performance, tailored to mitigate
extreme cold weather challenges for specific DOD applications, and tested based on
stakeholder needs and priorities.

256
Aircrew Labor In-Cockpit Automa-
tion System (ALIAS)
Dr. Stuart H. Young

https://www.darpa.mil/program/aircrew-labor-in-cockpit-automation-system.html

GPT-3 Summary: GOAL: The ALIAS program will develop a new system to
improve aircraft automation and reduce pilot workload.
TECHNOLOGY METHODS: Develop a new system that can be easily integrated
into existing aircraft. This system will be able to handle all aspects of a mission,
from takeoff to landing.
IMPACT: Improve flight safety and reduce pilot workload.

Full Details: Military aircraft have evolved to incorporate ever more automated
capabilities, improving mission safety and success rates. Yet operators of even the
most automated aircraft must still manage dauntingly complex interfaces and be
prepared to respond effectively in emergencies and other unexpected situations that
no amount of training can fully prepare one for. Avionics and software upgrades can
help, but the high cost of such improvements—which can run into the tens of millions
of dollars per aircraft—has limited the development, testing and fielding of novel
automation capabilities. To help overcome these challenges, DARPA created the
Aircrew Labor In-Cockpit Automation System (ALIAS) program. ALIAS envisions
a tailorable, drop-in, removable kit that would promote the addition of high levels of
automation into existing aircraft, enabling operation with reduced onboard crew. The
program intends to leverage the considerable advances made in aircraft automation
systems over the past 50 years, as well as similar advances in remotely piloted aircraft
automation, to help reduce pilot workload, augment mission performance and improve
aircraft safety. As an automation system, ALIAS aims to support execution of an
entire mission from takeoff to landing, even in the face of contingency events such as
aircraft system failures. ALIAS system attributes, such as persistent-state monitoring
and rapid recall of flight procedures, would further enhance flight safety. Easy-to-use
touch and voice interfaces would facilitate supervisor-ALIAS interaction. ALIAS
would also provide a platform for integrating additional automation or autonomy
capabilities tailored for specific missions.

257
Semantic Forensics (SemaFor)
Dr. William Corvey

https://www.darpa.mil/program/semantic-forensics.html

GPT-3 Summary: The Semantic Forensics (SemaFor) program seeks to develop

innovative semantic technologies for analyzing media. These technologies include se-
mantic detection algorithms, which will determine if multi-modal media assets have
been generated or manipulated. Attribution algorithms will infer if multi-modal me-
dia originates from a particular organization or individual. Characterization algo-
rithms will reason about whether multi-modal media was generated or manipulated
for malicious purposes. These SemaFor technologies will help identify, deter, and
understand adversary disinformation campaigns.

Full Details: Media generation and manipulation technologies are advancing rapidly
and purely statistical detection methods are quickly becoming insufficient for identi-
fying falsified media assets. Detection techniques that rely on statistical fingerprints
can often be fooled with limited additional resources (algorithm development, data,
or compute). However, existing automated media generation and manipulation algo-
rithms are heavily reliant on purely data driven approaches and are prone to making
semantic errors. For example, generative adversarial network (GAN)-generated faces
may have semantic inconsistencies such as mismatched earrings. These semantic
failures provide an opportunity for defenders to gain an asymmetric advantage. A
comprehensive suite of semantic inconsistency detectors would dramatically increase
the burden on media falsifiers, requiring the creators of falsified media to get ev-
ery semantic detail correct, while defenders only need to find one, or a very few,
inconsistencies.The Semantic Forensics (SemaFor) program seeks to develop innova-
tive semantic technologies for analyzing media. These technologies include semantic
detection algorithms, which will determine if multi-modal media assets have been
generated or manipulated. Attribution algorithms will infer if multi-modal media
originates from a particular organization or individual. Characterization algorithms
will reason about whether multi-modal media was generated or manipulated for ma-
licious purposes. These SemaFor technologies will help identify, deter, and under-
stand adversary disinformation campaigns.For more information, please see the Se-
maFor BAA.HR001119S0085: Proposers Day VideoHR001119S0085: Proposers Day
SlidesHR001119S0085: FAQ

258
System Security Integration Through
Hardware and Firmware (SSITH)
Dr. Lok Yan

https://www.darpa.mil/program/ssith.html

GPT-3 Summary: The SSITH program seeks to protect electronic systems from
common means of exploitation. Instead of relying on patches to ensure the safety of
software applications, SSITH aims to address the underlying hardware vulnerabilities
at the source. The program is developing hardware security architectures and associ-
ated design tools to protect systems against entire classes of vulnerabilities exploited
through software, not just specific vulnerability instances.

Full Details: Electronic systems have become a critical part of daily life. Due to in-
creased proliferation and reliance on these systems, their security is paramount to the
Department of Defense (DoD), commercial industry, and beyond. Current efforts to
protect electronic systems, however, rely on developing and deploying patches to the
software layer, without addressing underlying vulnerabilities in the hardware.The Sys-
tem Security Integration Through Hardware and Firmware (SSITH) program seeks to
protect electronic systems from common means of exploitation. Instead of relying on
patches to ensure the safety of software applications, SSITH aims to address the un-
derlying hardware vulnerabilities at the source. The program is developing hardware
security architectures and associated design tools to protect systems against entire
classes of vulnerabilities exploited through software, not just specific vulnerability in-
stances. SSITH is focusing specifically on common classes of hardware weaknesses as
identified by the MITRE Common Weakness Enumeration Specification (CWE) and
NIST, including buffer errors; information leakage; resource management; numeric
errors; injection; permissions, privileges, and access control; and hardware/system-
on-chip implementation errors. Researchers are exploring a number of different ap-
proaches that go well beyond patching. These include using metadata tagging to
detect unauthorized system access; utilizing context sensing pipelines to determine
the intent of instructions; and employing formal methods to reason about integrated
circuit systems and guarantee the accuracy of security characteristics.Throughout
the life of the program, SSITH seeks to continuously evaluate and strengthen the
secure hardware architectures in development. To support this evaluation, SSITH
is developing software tools to quantitatively measure the security properties of the
hardware architectures and to express and reason about security architectures at the
abstract (model) level and the concrete (product) level. Further, SSITH is exploring

259
external evaluation efforts that involve crowdsourced red teaming of the technologies
in development. This included DARPA’s first ever bug bounty program, Finding Ex-
ploits to Thwart Tampering (FETT) Bug Bounty. Additional information on FETT
is available here, https://fett.darpa.mil/To accomplish its goal, SSITH encourages
collaboration between academic, commercial, and the defense industrial base to pro-
vide robust and flexible solutions applicable to both DoD and commercial electronic
systems. SSITH has successfully developed a number of RISC-V FPGA-based tech-
nology demonstrations, and the technologies have been incorporated into commercial
designs. The ultimate goal of the program is to produce SSITH ASICs with near-term
applicability to a range of systems – from embedded systems found in mobile phones
and IoT devices to high performance servers in the cloud and military systems.

260
COnstructive Machine-learning Bat-
tles with Adversary Tactics (COM-
BAT)
Dr. Paul Zablocky

https://www.darpa.mil/program/constructive-machine-learning-battles-with-adversary-
tactics.html

GPT-3 Summary: GOAL: The COMBAT program will develop artificial intelli-
gence (AI) algorithms to generate models of Red Force brigade behaviors that chal-
lenge and adapt to Blue Forces in simulation experiments.
TECHNOLOGY METHODS: Deliver an AI Red Force capable of reacting in ways
not easily replicated by human role-players.
IMPACT: In future war games, COMBAT will deliver an AI Red Force capable of
reacting in ways not easily replicated by human role-players, challenging Blue role-
players and stimulating the development of new Blue tactics.

Full Details: Currently, Army force-on-force war games employ Red Team (adver-
sary) operators who rely upon long-established techniques, tactics, procedures, and
conventional force employment against their Blue opponent (U.S. Forces). COMBAT
aims to develop artificial intelligence (AI) algorithms to generate models of Red Force
brigade behaviors that challenge and adapt to Blue Forces in simulation experiments.
This effort will investigate how an adversary may evolve from current to future tac-
tics based on Blue Force actions and responses. If successful, COMBAT will rapidly
develop multiple feasible Red Force courses of action, identify optimal solutions, and
provide reasoning to support recommendations. In future war games, COMBAT will
deliver an AI Red Force capable of reacting in ways not easily replicated by human
role-players, challenging Blue role-players and stimulating the development of new
Blue tactics.

261
Explainable Artificial Intelligence (XAI)
Dr. Matt Turek

https://www.darpa.mil/program/explainable-artificial-intelligence.html

GPT-3 Summary: The goal of the XAI program is to create machine learning tech-
niques that can produce more explainable models, while still maintaining a high level
of learning performance. The methods used will be a combination of new and mod-
ified machine learning techniques, combined with state-of-the-art human-computer
interface techniques. The impact of this program will be the development of a suite
of machine learning techniques that will enable human users to understand, appropri-
ately trust, and effectively manage the emerging generation of artificially intelligent
machine partners.

Full Details: Figure 1. The Need for Explainable AI Dramatic success in machine
learning has led to a torrent of Artificial Intelligence (AI) applications. Continued ad-
vances promise to produce autonomous systems that will perceive, learn, decide, and
act on their own. However, the effectiveness of these systems is limited by the ma-
chine’s current inability to explain their decisions and actions to human users (Figure
1). The Department of Defense (DoD) is facing challenges that demand more intel-
ligent, autonomous, and symbiotic systems. Explainable AI—especially explainable
machine learning—will be essential if future warfighters are to understand, appropri-
ately trust, and effectively manage an emerging generation of artificially intelligent
machine partners. The Explainable AI (XAI) program aims to create a suite of ma-
chine learning techniques that: Produce more explainable models, while maintaining
a high level of learning performance (prediction accuracy); and Enable human users
to understand, appropriately trust, and effectively manage the emerging generation
of artificially intelligent partners. New machine-learning systems will have the ability
to explain their rationale, characterize their strengths and weaknesses, and convey
an understanding of how they will behave in the future. The strategy for achieving
that goal is to develop new or modified machine-learning techniques that will pro-
duce more explainable models. These models will be combined with state-of-the-art
human-computer interface techniques capable of translating models into understand-
able and useful explanation dialogues for the end user (Figure 2). Our strategy is to
pursue a variety of techniques in order to generate a portfolio of methods that will
provide future developers with a range of design options covering the performance-
versus-explainability trade space. Figure 2. XAI Concept XAI is one of a handful
of current DARPA programs expected to enable “third-wave AI systems”, where
machines understand the context and environment in which they operate, and over

262
time build underlying explanatory models that allow them to characterize real world
phenomena. The XAI program is focused on the development of multiple systems
by addressing challenge problems in two areas: (1) machine learning problems to
classify events of interest in heterogeneous, multimedia data; and (2) machine learn-
ing problems to construct decision policies for an autonomous system to perform a
variety of simulated missions. These two challenge problem areas were chosen to
represent the intersection of two important machine learning approaches (classifica-
tion and reinforcement learning) and two important operational problem areas for
the DoD (intelligence analysis and autonomous systems). In addition, researchers are
examining the psychology of explanation. XAI research prototypes are tested and
continually evaluated throughout the course of the program. In May 2018, XAI re-
searchers demonstrated initial implementations of their explainable learning systems
and presented results of initial pilot studies of their Phase 1 evaluations. Full Phase
1 system evaluations are expected in November 2018. At the end of the program,
the final delivery will be a toolkit library consisting of machine learning and human-
computer interface software modules that could be used to develop future explainable
AI systems. After the program is complete, these toolkits would be available for fur-
ther refinement and transition into defense or commercial applications.

263
Domain-Specific System on Chip (DSSoC)
Dr. Thomas Kazior

https://www.darpa.mil/program/domain-specific-system-on-chip.html

GPT-3 Summary: GOAL: The Domain-Specific System on Chip (DSSoC)

program seeks to develop a heterogeneous system-on-chip (SoC) comprised of many
cores that mix general purpose processors, special purpose processors, hardware
accelerators, memory, and input/output (I/O) devices to significantly improve
performance of applications within a domain.

METHODS: DSSoC will address critical issues in today’s development cycle

that force low-level engineering to port applications onto the specifics of the
underlying processors. Specifically, the DSSoC program will seek to develop software
stacks that enable better hardware

Full Details: The general-purpose computer has remained the dominant computing
architecture for the last 50 years, driven largely by the relentless pace of Moore’s Law.
As this trajectory shows signs of slowing, however, it has become increasingly more
challenging to achieve performance gains from generalized hardware, setting the stage
for a resurgence in specialized architectures. Today’s specialized, application-specific
integrated circuits (ASICs) — hardware customized for a specific application — offer
limited flexibility and are costly to design, fabricate, and program. The Domain-
Specific System on Chip (DSSoC) program seeks to prove that there need not be a
continued tradeoff between efficiency, like that found in ASICs, and flexibility, the
hallmark of general-purpose processors. The goal of DSSoC is to develop a heteroge-
neous system-on-chip (SoC) comprised of many cores that mix general purpose pro-
cessors, special purpose processors, hardware accelerators, memory, and input/output
(I/O) devices to significantly improve performance of applications within a domain. A
domain is larger than any one application, where one processor can effectively address
problems more efficiently than a general purpose processor but without the challenge,
time, and cost of building a special-purpose system like an ASIC. DSSoC is exploring
architectures that improve the efficiency of computing through specialized processing
while maintaining programmability. To better support the large and growing range
of applications needed by the Department of Defense (DoD) to maintain its tech-
nological advantage the DSSoC program will seek to produce concepts that provide
improved computing efficiency for embedded processing needs while making these
systems more programmable. To do this, DSSoC will address critical issues in to-
day’s development cycle that force low-level engineering to port applications onto the

264
specifics of the underlying processors. Specifically, the DSSoC program will seek to
develop software stacks that enable better hardware-software co-design by providing
vertical integration of tools from the hardware up to the development environment.
A second key impact of DSSoC will be the development of intelligent scheduling of
applications and data on a heterogeneous processor to make better use of the hetero-
geneity and parallelism of these systems, whereas today, programming applications
for complex systems is a time-consuming, hand-tuned exercise.

265
Make-It
Dr. Vishnu Sundaresan

https://www.darpa.mil/program/make-it.html

GPT-3 Summary: The goal of the Make-It program is to automate small molecule
discovery and synthesis in order to create new materials faster and more efficiently.
The technology methods used include artificial intelligence to plan and optimize syn-
thetic routes, coupled with fully automated synthesis that includes algorithms for au-
tomation and process control, interconnected fluidic modules for continuous synthesis,
and in-line characterization and purification. The expected impact is the transforma-
tion of synthetic chemistry into an information-centric science, accelerating the pace
of chemical innovation and small molecule manufacturing.

Full Details: Synthetic chemistry is important across countless technological areas,

from medicines to energetics to advanced coatings to functional materials. While our
synthetic capabilities have developed rapidly over the last century, current approaches
are still slow and inefficient, with poor reproducibility and scalability and limited use
of prior knowledge. Such an approach not only limits production of known materials,
but also impedes discovery of better synthetic routes and completely new molecules.
The DARPA Make-It program is automating small molecule discovery and synthesis
to propel the field beyond conventional batch-based, intuition-driven capabilities.
Make-It is developing artificial intelligence-based approaches to plan and optimize
synthetic routes, coupled with methods for fully automated synthesis that include
algorithms for automation and process control, interconnected fluidic modules for
continuous synthesis, and in-line characterization and purification. Researchers are
also working on methods to rapidly explore the vast parameter space associated with
synthesis, which has only been minimally sampled by hand thus far. Make-It seeks
to provide the foundational technologies required to transform synthetic chemistry
to an information-centric science, accelerating the pace of chemical innovation and
small molecule manufacturing.

266
Driven and Nonequilibrium Quan-
tum Systems (DRINQS)
Dr. Tatjana Curcic

https://www.darpa.mil/program/driven-and-nonequilibrium-quantum-systems.html

GPT-3 Summary: GOAL: The DRINQS program will investigate whether driving
a quantum system out of equilibrium can increase the length of time that its quantum
state endures.

TECHNOLOGY METHODS: The program will study what protocols can op-
timally enhance quantum coherence in driven systems.

IMPACT: If successful, the program could lead to mobile high-performance

atomic clocks that could provide troops GPS-like precision navigation in GPS-denied
environments.

Full Details: DRINQS is a fundamental science program that aims to investigate a

recent paradigm shift in quantum research, which maintains that periodically driving
a system out of equilibrium may increase the length of time that its quantum state
endures. DRINQS aims to investigate this phenomenon and demonstrate significant
gains over conventional states in timekeeping, field sensing, and information process-
ing for use in national security applications. The performance of quantum sensors
and devices is intimately dependent on its coherence, which is the length of time the
underlying system retains its quantum properties before environmental interactions
make the state behave like a conventional classical system. The interactions of atoms
within the system and with a “noisy” thermal environment are typically limiting fac-
tors of coherence time, which means the best quantum devices require extremely clean
control signals and cryogenic environments to reduce thermal disturbances. This re-
quirement for pristine laboratory conditions to maintain coherence has limited the
applicability and adoption of quantum technology in operational devices of interest to
national security, such as mobile high-performance atomic clocks that could provide
troops GPS-like precision navigation in GPS-denied environments. DRINQS aims to
determine and demonstrate what protocols can optimally enhance quantum coher-
ence in driven systems. Additionally, the program plans to demonstrate a proof-of-
principle with 10X improvement over conventional techniques in clock stability, high
spatial resolution field sensing, and quantum information applications.

267
Optomechanical Thermal Imaging (Op-
TIm)
Dr. Mukund Vengalattore

https://www.darpa.mil/program/optomechanical-thermal-imaging.html

GPT-3 Summary: The goal of the OpTIm program is to develop room-temperature

IR detectors with quantum-limited sensitivity. The methods used will include op-
tomechanical sensitivity, quantum-limited all-optical readout, and dynamically tun-
able narrowband or multispectral detection capabilities. The impact of this program
will be the establishment of the fundamental performance limits of this new modality
of infrared detection.

Full Details: Infrared (IR) detectors underpin a vast application space including
night vision, battlefield surveillance, terrestrial and space imaging, biochemical fin-
gerprinting, and non-invasive medical diagnosis. Current IR detectors fall in one of
two categories – thermal sensors that detect IR radiation as heat, and photodetectors
that convert IR radiation into electric current. While the former class of sensors can
operate at room temperature, they are highly limited in their sensitivity and response
time. In contrast, while IR photodetectors are capable of quantum-limited sensitiv-
ity, they require cryogenic cooling, which limits their widespread use. The OpTIm
program seeks to develop integrated, room-temperature, and quantum-limited IR
detectors that bridge the current technological gap between low performance, room-
temperature bolometers and cryogenically-cooled photodetectors. Performers in Op-
TIm will seek to develop, validate and benchmark a new class of room-temperature
IR detectors that combine optomechanical sensitivity, quantum-limited all-optical
readout, and dynamically tunable narrowband or multispectral detection capabili-
ties. OpTIm also aims to establish the fundamental performance limits of this new
modality of infrared detection.

268
Ocean of Things
Mr. John Waterston

https://www.darpa.mil/program/ocean-of-things.html

GPT-3 Summary: The Ocean of Things program seeks to create a network of small,
low-cost floats that can collect data about the environment and activity in the ocean.
The program’s challenges include designing floats that can survive in harsh conditions
and developing software to process and analyze the data collected by the floats. If
successful, the program could provide persistent maritime situational awareness and
help to improve oceanographic and meteorological models.

Full Details: DARPA’s Oceans of Things program seeks to enable persistent mar-
itime situational awareness over large ocean areas by deploying thousands of small,
low-cost floats that form a distributed sensor network. Each smart float contains
a suite of commercially available sensors to collect environmental data-such as sea
surface temperature, sea state, and location - as well as activity data about com-
mercial vessels, aircraft, and even maritime mammals moving through the area. The
floats transmit data periodically via satellite to a cloud network for storage and real-
time analysis.The technical challenge for Ocean of Things lies in two key areas: float
development and data analytics.Under float development, performers are tasked to
design an intelligent float to house a passive sensor suite that can survive in harsh
maritime environments. Each float processes information on the edge before report-
ing it, persisting for up to one year before safely scuttling itself in the deep ocean.
The floats are required to be made of environmentally safe materials, pose no dan-
ger to vessels, and comply with all federal laws, regulations, and executive orders
related to protection of marine life.The data analytics portion of the Ocean of Things
program requires performers to develop cloud-based software and advanced analytic
techniques to process the floats’ reported data. This effort includes dynamic display
of float locations, health, and mission performance; processing of environmental data
for oceanographic and meteorological models; developing algorithms to automatically
detect, track, and identify nearby vessels; and identification of new indicators of mar-
itime activity.For the latest on Ocean of Things including float data made available
for the wider oceanographic research community, visit oceanofthings.darpa.mil.

269
Grounded Artificial Intelligence Lan-
guage Acquisition (GAILA)
Dr. William Corvey

https://www.darpa.mil/program/grounded-artificial-intelligence-language-acquisition.html

GPT-3 Summary: The GAILA program seeks to research and develop a model for
grounded language acquisition and an automated language acquisition prototype that
learns to understand English text and speech, making the information more useable
by automated analytics.

Full Details: The past few decades have seen explosive growth in the development
and training of artificial intelligence (AI) systems, which are now embedded in digital
computing processes spanning several key industries. One area that has benefited
from AI, and specifically machine learning (ML) techniques and statistical methods,
is the area of human language technology (HLT). These methods have provided signif-
icant improvements to numerous language technologies such as machine translation,
information retrieval, name entity detection, event detection, knowledge base cre-
ation, and more. These technologies have been successfully deployed in government,
defense, and commercial settings, such as for call centers, reservation services, and
personal assistants. However, ML suffers from a need for large amounts of annotated
data for training in order to achieve the required accuracy for various applications.
ML technology is also brittle, incapable of dealing with new data sources, topics,
media, and vocabularies. These weaknesses of ML, as applied to natural language,
are due to exclusive reliance on the statistical aspects of language, with no regard for
its meaning. The Grounded Artificial Intelligence Language Acquisition (GAILA) AI
Exploration (AIE) topic aims to enable computers to acquire language in a manner
similar to the way children do. Children acquire language based on their perceptions
of aural and visual information about the world around them. The process of observ-
ing moving images and aligning them with auditory stimuli allows them to associate
a sequence of sounds (a word) with some aspect of the concrete or abstract elements
of the world that the word represents. Children learn to decipher which aspects of
an observed scenario relate to the different words in the message from a tiny frac-
tion of the examples that ML systems require. Sequencing information, variations
of word forms, and other additional information help children make ever finer clas-
sifications of the concepts that they learn, about events or actions (often expressed
by verbs), the objects or entities that participate in those events (typically nouns),
and the attributes and relations of those entities and events (adjectives and preposi-

270
tional phrases). GAILA seeks to research and develop a model for grounded language
acquisition and an automated language acquisition prototype that learns to under-
stand English text and speech, making the information more useable by automated
analytics.

271
Cross-Domain Maritime Surveillance
and Targeting (CDMaST)
Mr. Aaron Kofford

https://www.darpa.mil/program/cross-domain-maritime-surveillance-and-targeting.html

GPT-3 Summary: GOAL: The Cross-Domain Maritime Surveillance and Target-

ing (CDMaST) program seeks to identify and implement architectures consisting of
novel combinations of manned and unmanned systems to deny ocean environments
to adversaries as a means of projecting power.

TECHNOLOGY METHODS: By exploring promising new developments in

distributed unmanned systems, advanced sensors, resilient communications and
emerging long-range weapon systems, the program aims to develop an advanced,
integrated undersea warfighting capability able to hold adversary submarines at risk
over large contested maritime areas.

IMPACT: The desired result is to both deliver an experimentally validated

effective

Full Details: The Cross-Domain Maritime Surveillance and Targeting (CDMaST)

program seeks to identify and implement architectures consisting of novel combina-
tions of manned and unmanned systems to deny ocean environments to adversaries as
a means of projecting power. By exploring promising new developments in distributed
unmanned systems, advanced sensors, resilient communications and emerging long-
range weapon systems, the program aims to develop an advanced, integrated undersea
warfighting capability able to hold adversary submarines at risk over large contested
maritime areas. The program encompasses both development of warfighting architec-
tures and underlying technologies as well as experimentation to explore and test the
viability and effectiveness of these complex architectures. A key component is the
development of the Distributed Experimentation Environment (DE2), a government
owned and operated constructive, virtual, and live testbed and system of systems
technology integration process. The DE2 process integrates and evolves novel tech-
nologies through a phased development strategy consisting of architecture refinement,
engineering testing, and operational experiments in operationally relevant contexts.
The desired result is to both deliver an experimentally validated effective undersea
warfighting architecture; and provide an enduring capability for the Department of
Defense to efficiently evolve, innovate, and deliver effective warfighting architectures.

272
Reversible Quantum Machine Learn-
ing and Simulation (RQMLS)
Dr. Joe Altepeter

https://www.darpa.mil/program/reversible-quantum-machine-learning-and-simulation.html

GPT-3 Summary: GOAL: The RQMLS program will explore the potential of
quantum annealers for machine learning tasks, with the goal of developing a scalable
system.
TECHNOLOGY METHODS: The program will test predictions on small-scale quan-
tum annealers to determine their potential for machine learning tasks.
IMPACT: If successful, the program could develop a scalable quantum annealer sys-
tem with the potential to transform machine learning.

Full Details: Machine learning and artificial intelligence techniques are currently
being applied in a diverse number of fields, including molecular simulation, many-
body physics, classification, and computational optimization. However, progress in
addressing these types of problems is being slowed or stopped when the problem com-
plexity grows exponentially with problem size. Moreover, even when these complexity
barriers are overcome, the impact of machine learning solutions are often mitigated
by the high energy cost of training and operating the machine learning systems.In
principle, both of these fundamental obstacles – exponentially growing complexity
and energy inefficiency – might be overcome using high-coherence quantum anneal-
ers, which are a specific type of quantum computing technology.DARPA’s Reversible
Quantum Machine Learning and Simulation (RQMLS) AIE opportunity aims to: (1)
explore the fundamental limits of reversible quantum annealers; (2) quantitatively
predict the computational utility of these systems for machine learning tasks in sim-
ulation, many-body physics, classification, optimization, and other fields; and (3)
design experimental tests for these predictions that can be carried out on small-scale
systems. If successful, these small-scale systems could be scaled to much larger, po-
tentially transformative systems.

273
Assured Micropatching (AMP)
Dr. Sergey Bratus

https://www.darpa.mil/program/assured-micropatching.html

GPT-3 Summary: The goal of the Assured Micropatching program is to create the
capability for rapid patching of legacy binaries in mission critical systems, even when
the original source code version and/or build process is not available. The program
seeks to address gaps in the current software development paradigm through break-
throughs in and novel approaches to technical challenges, including but not limited to:

- Identifying modular units in executable binary images, and identifying mod-

ules’ interfaces, interactions, and linking artifacts to enable subsequent assured
relinking and re-integration of patched binary modules
- Decompiling the executable binary code into forms suitable

Full Details: Our society’s infrastructure is increasingly dependent on software de-

ployed on a wide variety of computing devices other than commodity personal com-
puters, such as industrial equipment, automobiles, and airplanes. Unlike commodity
computers that have short upgrade cycles and are easily replaceable in case of fail-
ure, these computing devices are intended for longer service, and are hard to replace.
Thus, the amount of deployed software that needs to be maintained is continually in-
creasing, while the growing use of telemetry on such devices potentially exposes their
software to cyber-attacks. To fix cybersecurity flaws in software, vendors distribute
patched versions of the software. Unfortunately, even after a particular flaw has been
fully understood, and a remediation approach has been developed and expressed as
a source code change in the current version of the software, the ability of vendors
to produce patches for all of their deployed devices in a timely, assuredly safe, and
scalable manner is limited. Additional challenges arise when the exact source code
version has been lost, the process for building the software from source code was not
documented, and/or the original software development environment is not available.
These limitations and challenges result in mission-critical software going unpatched
for months to years, increasing the opportunity for attackers. The goal of the Assured
Micropatching (AMP) program is to create the capability for rapid patching of legacy
binaries in mission critical systems, including the cases where the original source code
version and/or build process is not available. AMP aims to create new capabilities to
analyze, modify, and fix legacy software in binary form, capable of producing assured
targeted micropatches for known security flaws in existing binaries. Micropatches
change the fewest possible bytes to achieve their objective, which minimizes poten-

274
tial side effects, and should enable proofs that the patches will preserve the original
baseline functionality of the system. With these proofs, the time to test and deploy
the patched system should be reduced from months to days.To achieve this goal, the
AMP program seeks to address gaps in the current software development paradigm
through breakthroughs in and novel approaches to technical challenges, including but
not limited to:Identifying modular units in executable binary images, and identifying
modules’ interfaces, interactions, and linking artifacts to enable subsequent assured
relinking and re-integration of patched binary modules;Decompiling the executable
binary code into forms suitable for automatically situating a patch for a known se-
curity flaw existing in the binary;Generating minimal-change binary micropatches
for existing binary images and for rigorous reasoning about their effects and testing
these effects to verify noninterference of the changes with the binary’s baseline func-
tionality; andUsing available sources of information, such as source code and binary
samples, to recover missing relevant parts of the source code and the build process.

275
Signature Management using Oper-
ational Knowledge and Environments
(SMOKE)
Mr. Tejas Patel

https://www.darpa.mil/program/signature-management-using-operational-knowledge-
and-environments.html

GPT-3 Summary: The goal of the SMOKE program is to develop data-driven

tools to automate the planning and execution of threat-emulated cyber infrastruc-
ture needed for network security assessments. The methods of the program include
developing data-driven tools to automate the discovery of distinguishable patterns of
sophisticated cyber threat infrastructure. The impact of the program is to prototype
components that enable red teams to plan, build, and deploy cyber infrastructure
that is informed by machine-readable signatures of sophisticated cyber threats.

Full Details: Networks are under persistent threat from malicious cyber actors
(MCAs). In response, a growing industry of network security professionals are of-
fering realistic, threat-informed assessments of network owners’ defensive posture.
These assessments are performed by a team of ethical hackers (i.e., the red team)
in which they assume the role of sophisticated MCAs and perform a controlled se-
curity test in collaboration with network defenders (i.e., the blue team). Red team
exercises are designed to exceed simple penetration testing and emulate MCA be-
haviors as realistically as possible. Realistic emulation of sophisticated cyber threats
in a measured exercise is very helpful for providing a comprehensive picture of net-
work defenders’ readiness.The Signature Management using Operational Knowledge
and Environments (SMOKE) program aims to develop data-driven tools to automate
the planning and execution of threat-emulated cyber infrastructure needed for net-
work security assessments (e.g., red team exercises). In a complementary activity,
SMOKE will explore the development of data-driven tools to automate the discovery
of distinguishable patterns of sophisticated cyber threat infrastructure (i.e., signa-
tures). Together, SMOKE seeks to prototype components that enable red teams to
plan, build, and deploy cyber infrastructure that is informed by machine-readable
signatures of sophisticated cyber threats. Additional information is available in the
SMOKE Broad Agency Announcement

276
Arcadia
Dr. Paul Sheehan

https://www.darpa.mil/program/arcadia.html

GPT-3 Summary: GOAL: Develop ”probiotics” for military materiel that use
naturally occurring microorganisms to build protective coatings.
TECHNOLOGY METHODS: Use microorganisms to generate coatings that prevent
corrosion, decrease drag, or inhibit growth of black mold.
IMPACT: Improve the performance and lifespan of DoD assets.

Full Details: Bacteria are the most abundant and diverse lifeform on Earth, with
the majority spending their lives in biofilms, a thin layer of microorganisms (e.g. bac-
teria) that can coat or form on various surfaces. This lifestyle is universally perceived
as problematic since biofilms contribute significantly to equipment degradation, in-
cluding Department of Defense (DoD) assets. However, biofilms do not have to be
a problem. New insights suggest biofilms could be transfigured by redirecting their
composition and structure. Similar to the Arcadian vision of harmony with nature,
the Arcadia program aims to develop “probiotics” for military materiel using mi-
croorganisms that naturally occur on DoD assets to build protective communities.
These probiotics will generate robust and beneficial coatings that prevent corrosion,
decrease drag, or inhibit the growth of black mold on DoD assets.

277
Social Simulation for Evaluating On-
line Messaging Campaigns (Social-
Sim)
Dr. Brian Kettler

https://www.darpa.mil/program/computational-simulation-of-online-social-behavior.html

GPT-3 Summary: The goal of the SocialSim program is to develop innovative

technologies for high-fidelity computational simulation of online social behavior. The
focus will be on information spread and evolution. SocialSim will aim to develop
novel approaches to address the challenges of accuracy and scalability in social and
behavioral simulation.

Full Details: A rapidly increasing percentage of the world’s population is connected

to the global information environment. At the same time, the information environ-
ment is enabling social interactions that are radically changing how and at what rate
information spreads. Both nation-states and nonstate actors have increasingly drawn
upon this global information environment to promote their beliefs and further related
goals. A simulation of the spread and evolution of online information, if accurate and
at-scale, could enable a deeper and more quantitative understanding of adversaries’
use of the global information environment than is currently possible using existing
approaches. At present, the U.S. Government employs small teams of experts to
speculate how information may spread online. While these activities provide some
insight, they take considerable time to orchestrate and execute, the accuracy with
which they represent real-world online behavior is unknown, and their scale (in terms
of the size and granularity with which populations are represented) is such that they
can represent only a fraction of the real world. High-fidelity (i.e., accurate, at-scale)
computational simulation of the spread and evolution of online information would
support efforts to analyze strategic disinformation campaigns by adversaries, deliver
critical information to local populations during disaster relief operations, and could
potentially contribute to other critical missions in the online information domain.
The goal of Computational Simulation of Online Social Behavior (SocialSim) is to
develop innovative technologies for high-fidelity computational simulation of online
social behavior. SocialSim will focus specifically on information spread and evolution.
Current computational approaches to social and behavioral simulation are limited in
this regard. Top-down simulation approaches focus on the dynamics of a population
as a whole, and model behavioral phenomena by assuming uniform or mostly-uniform

278
behavior across that population. Such methods can easily scale to simulate massive
populations, but can be inaccurate if there are specific, distinct variations in the
characteristics of the population. In contrast, bottom-up simulation approaches treat
population dynamics as an emergent property of the activities and interactions taking
place within a diverse population. While such approaches can enable more accurate
simulation of information spread, they do not readily scale to represent large popu-
lations. SocialSim aims to develop novel approaches to address these challenges.

279
Learning with Less Labeling (LwLL)
Dr. Bruce Draper

https://www.darpa.mil/program/learning-with-less-labeling.html

GPT-3 Summary: GOAL: The Learning with Less Labeling (LwLL) program
aims to reduce the amount of labeled data required to train machine learning models
by up to six orders of magnitude.

METHODS: The LwLL program will focus on developing learning algorithms

that learn and adapt efficiently, and on formally characterizing machine learning
problems to prove the limits of learning and adaptation.

IMPACT: The LwLL program has the potential to reduce the cost and time
needed to train machine learning models, as well as the amount of data needed to
adapt models to new environments.

Full Details: In supervised machine learning (ML), the ML system learns by exam-
ple to recognize things, such as objects in images or speech. Humans provide these
examples to ML systems during their training in the form of labeled data. With
enough labeled data, we can generally build accurate pattern recognition models. The
problem is that training accurate models currently requires lots of labeled data. For
tasks like machine translation, speech recognition or object recognition, deep neural
networks (DNNs) have emerged as the state of the art, due to the superior accuracy
they can achieve. To gain this advantage over other techniques, however, DNN models
need more data, typically requiring 109 or 1010 labeled training examples to achieve
good performance. The commercial world has harvested and created large sets of
labeled data for training models. These datasets are often created via crowdsourcing:
a cheap and efficient way to create labeled data. Unfortunately, crowdsourcing tech-
niques are often not possible for proprietary or sensitive data. Creating data sets for
these sorts of problems can result in 100x higher costs and 50x longer time to label. To
make matters worse, machine learning models are brittle, in that their performance
can degrade severely with small changes in their operating environment. For instance,
the performance of computer vision systems degrades when data is collected from a
new sensor and new collection viewpoints. Similarly, dialog and text understanding
systems are very sensitive to changes in formality and register. As a result, additional
labels are needed after initial training to adapt these models to new environments and
data collection conditions. For many problems, the labeled data required to adapt
models to new environments approaches the amount required to train a new model

280
from scratch. The Learning with Less Labeling (LwLL) program aims to make the
process of training machine learning models more efficient by reducing the amount of
labeled data required to build a model by six or more orders of magnitude, and by
reducing the amount of data needed to adapt models to new environments to tens to
hundreds of labeled examples. In order to achieve the massive reductions of labeled
data needed to train accurate models, the LwLL program will focus on two technical
objectives: Development of learning algorithms that learn and adapt efficiently; and
Formally characterize machine learning problems and prove the limits of learning and
adaptation. Additional information is available in the LwLL BAA

281
Robotic Servicing of Geosynchronous
Satellites (RSGS)
Ms. Ana Saplan

https://www.darpa.mil/program/robotic-servicing-of-geosynchronous-satellites.html

GPT-3 Summary: GOAL: The RSGS program will develop technologies to en-
able cooperative inspection and servicing of satellites in geosynchronous Earth orbit
(GEO).
TECHNOLOGY METHODS: Develop a modular toolkit of hardware and software
to be joined with a privately developed spacecraft. Test these technologies on orbit
within the next five years.
IMPACT: The RSGS program could enable commercial spacecraft to make house
calls in space, providing inspection, diagnosis, and repair services for satellites.

Full Details: Hundreds of military, government and commercial satellites reside

today in geosynchronous Earth orbit (GEO) some 22,000 miles (36,000 kilometers)
above the Earth—a perch ideal for providing communications, meteorology and na-
tional security services, but one so remote as to preclude inspection and diagnosis
of malfunctioning components, much less upgrades or repairs. Even fully functional
satellites sometimes find their working lives cut short simply because they carry obso-
lete payloads—a frustrating situation for owners of assets worth hundreds of millions
of dollars. With no prospects for assistance once in orbit, satellites destined for
GEO today are loaded with backup systems and as much fuel as can be accommo-
dated, adding to their complexity, weight and cost. But what if help was just a
service call away?DARPA’s Robotic Servicing of Geosynchronous Satellites (RSGS)
program intends to answer that question by developing technologies that would enable
cooperative inspection and servicing in GEO and demonstrating those technologies
on orbit within the next five years. Under the RSGS vision, a DARPA-developed
modular toolkit, including hardware and software, would be joined to a privately
developed spacecraft to create a commercially owned and operated robotic servicing
vehicle (RSV) that could make house calls in space. By executing the RSGS program,
DARPA seeks to:Demonstrate in or near GEO that a robotic servicing vehicle can
perform safe, reliable, useful and efficient operations, with the flexibility to adapt to
a variety of on-orbit missions and conditions Demonstrate satellite servicing mission
operations on operational GEO satellites in collaboration with commercial and U.S.
Government spacecraft operators Support the development of a servicer spacecraft
with sufficient propellant and payload robustness to enable dozens of missions over

282
several years

283
Robust Optical Clock Network (ROCkN)
Dr. Tatjana Curcic

https://www.darpa.mil/program/robust-optical-clock-network.html

GPT-3 Summary: GOAL: The ROCkN program seeks to develop optical precision
timing technologies to significantly increase the precision and holdover of DoD timing
capability in low-size, weight, and power (SWaP) devices that can operate outside
the lab.

TECHNOLOGY METHODS: TA1 aims to develop small portable clocks that

provide the time precision necessary for distributed coherent sensing and other
applications in a robust package suitable for airborne or spaceborne platforms.
TA2 aims to develop transportable clocks with month-long holdover of GPS-quality
precise time and autonomous operation, with a view toward providing a

Full Details: Over the past two decades, precision timing in the optical domain
has advanced rapidly in the laboratory and has recently surpassed – in stability and
precision – the ubiquitous microwave timing techniques that make up the foundations
of global precision timing in commercial and Department of Defense (DoD) domains.
Optical precision timing techniques provide a means for orders of magnitude higher
precision and accuracy, but, just as importantly, they enable more resilient timing
capabilities with less reliance on GPS by virtue of longer holdover times and usage
of optical signals that are more difficult to jam or spoof.The “Robust Optical Clock
Network” (ROCkN) program seeks to develop optical precision timing technologies
to significantly increase the precision and holdover of DoD timing capability in low-
size, weight, and power (SWaP) devices that can operate outside the lab. ROCkN
has two Technical Areas (TA) based on different application domains. TA1 aims to
develop small portable clocks that provide the time precision necessary for distributed
coherent sensing and other applications in a robust package suitable for airborne
or spaceborne platforms. TA2 aims to develop transportable clocks with month-
long holdover of GPS-quality precise time and autonomous operation, with a view
toward providing a local regional time scale on land or seaborne platforms. The
common technical challenge is to provide the timing precision and holdover required
while constrained by SWaP and environmental sensitivity requirements.If successful,
ROCkN will transition recent advances in laboratory optical timing technologies out
of the lab and toward devices operational in the field to enable order of magnitude
improvements in DoD timing capabilities.

284
Focal arrays for Curved Infrared Im-
agers (FOCII)
Dr. Dev Palmer

https://www.darpa.mil/program/focal-arrays-for-curved-infrared-imagers.html

GPT-3 Summary: The goal of the FOCII program is to develop and demonstrate
scalable technologies for curving large format, high performance infrared focal plane
arrays to a small radius of curvature to maximize performance. The program also
seeks to demonstrate curving of smaller format focal plane arrays to an extreme radius
of curvature to enable the smallest form factors possible while maintaining exquisite
performance.

Full Details: Nearly all military-fielded imaging systems contain detector arrays fab-
ricated using planar processes developed for electronic integrated circuits, resulting
in the need to use large and complex optics to properly bring imaged objects into fo-
cus. Recent industry-led advancements have resulted in modestly curved, small-area,
visible-light image sensors. The Focal arrays for Curved Infrared Imagers (FOCII)
program plans to build upon the visible sensor advancements from industry by extend-
ing this capability towards large format cryogenically cooled infrared imagers with
extreme curvatures to vastly improve performance while reducing the weight, volume,
and cost of optics.While curved imagers have been attempted by the community in
the past, the approaches used were not able to produce commercially scalable, large
format devices with small radius of curvatures while maintaining high performance.
FOCII seeks to develop and demonstrate scalable technologies for curving existing
large format, high performance infrared focal plane arrays to a small radius of curva-
ture to maximize performance as well as demonstrate curving of smaller format focal
plane arrays to an extreme radius of curvature to enable the smallest form factors
possible while maintaining exquisite performance.

285
Recycling at the Point of Disposal
(RPOD)
Dr. Vishnu Sundaresan

https://www.darpa.mil/program/recycling-at-the-point-of-disposal.html

GPT-3 Summary: GOAL: Develop a way to recycle critical elements from elec-
tronic waste so that they can be reused.
METHODS: Develop novel extraction chemistries to recover critical elements from
e-waste.
IMPACT: Reduce the need for supply chain disruptions of critical elements sourced
or processed abroad.

Full Details: The Recycling at the Point of Disposal (RPOD) Disruption Opportu-
nity (DO) will evaluate the technical feasibility to recover (separate and coextract)
multiple low-volume fraction critical elements present in end-of-life electronic hard-
ware (e-waste). Separation is defined as the extraction of various elements sequen-
tially, and coextraction is defined as the extraction of a specified list of elements
simultaneously from a feedstock containing a mixture of elements with other con-
stituents. Contemporary approaches such as pyrometallurgy and hydrometallurgy
for recovering critical elements from e-waste are best suited to recover a large volume
fraction component from the feedstock and are ill-suited to recover multiple low-
volume fraction elements typically found in e-waste. While it is theoretically possible
to exploit a combination of physical, chemical, and functional properties for fractional
recovery of critical elements, such approaches are limited by insufficient knowledge on
chemistries that can work together at the point of disposal. DARPA’s Recycling at
the Point of Disposal (RPOD) program aims to develop novel extraction chemistries
and explore practical limits of yield, extraction efficiency, and purity to recover a
selective list of critical elements from commercial and DoD e-waste, anddemonstrate
the technology in a benchtop hardware prototype. If successful, the recovered stream
from e-waste can minimize supply chain disruptions of critical elements sourced or
processed abroad that are essential for high performance DoD hardware.

286
System-of-Systems Enhanced Small
Unit (SESU)
Dr. Paul Zablocky

https://www.darpa.mil/program/system-of-systems-enhanced-small-unit.html

GPT-3 Summary: GOAL: The SESU program will focus on two technical areas:
1) Adaptive command and control (C2) that enables distributed mission planning
and counter-A2/AD missions leveraging a large number of low-cost, autonomous,
unmanned air and ground platforms equipped with appropriate sensors and effectors;
and 2) Innovative sensors and effectors that can cooperate to destroy, disrupt, degrade,
and/or delay aspects of an adversary’s A2/AD capabilities.
TECHNOLOGY METHODS: Develop systems of autonomous platforms and sensors
that can work together to achieve objectives.
IM

Full Details: Future U.S. land forces are increasingly likely to face an adversary
force that is overwhelmingly superior in size and armament with formidable anti-
access/area denial (A2/AD) capabilities. SESU seeks to deliver system-of-systems
(SoS) capabilities that could enable a small unit (˜200-300 soldiers, corresponding
materiel footprint, and limited rear-echelon support) to destroy, disrupt, degrade,
and/or delay the adversary’s A2/AD and maneuver capabilities in order to enable
joint and coalition multi-domain operations at appropriate times and locations. The
SESU program will focus on two technical areas: 1) Adaptive command and control
(C2) that enables distributed mission planning and counter-A2/AD missions lever-
aging a large number of low-cost, autonomous, unmanned air and ground platforms
equipped with appropriate sensors and effectors; and 2) Innovative sensors and ef-
fectors that can cooperate to destroy, disrupt, degrade, and/or delay aspects of an
adversary’s A2/AD capabilities.

287
SIGMA+
Dr. Mark Wrobel

https://www.darpa.mil/program/sigma-plus.html

GPT-3 Summary: GOAL: The SIGMA+ program will expand SIGMA’s ability
to detect illicit nuclear and radioactive materials, as well as chemical, biological, and
explosives threats.
TECHNOLOGY METHODS: Develop new sensors and networks that would alert
authorities to various substances related to weapons of mass destruction.
IMPACT: Create a scalable CBRNE detection network that can cover a major
metropolitan city and its surrounding region.

Full Details: The SIGMA+ program aims to expand SIGMA’s advance capabil-
ity to detect illicit radioactive and nuclear materials by developing new sensors and
networks that would alert authorities to chemical, biological, and explosives threats
as well. SIGMA+ calls for the development of highly sensitive detectors and ad-
vanced intelligence analytics to detect minute traces of various substances related to
weapons of mass destruction (WMD) threats. SIGMA+ will use a common network
infrastructure and mobile sensing strategy, a concept that was proven effective in
the SIGMA program. The SIGMA+ chemical, biological, radiological, nuclear and
high-yield explosive (CBRNE) detection network would be scalable to cover a major
metropolitan city and its surrounding region. Planned execution of SIGMA+ will
occur in two phases. Phase 1 will focus on developing novel sensors for chemicals,
explosives, and biological agents while Phase 2 will focus on network development,
analytics and integration.

288