International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 07 Issue: 08 | Aug 2020 www.irjet.net p-ISSN: 2395-0072

A Study on Artificial Intelligence

N Satya Abhisht1
1III.B.TECH-AERO, GITAM UNIVERSITY, Rudraram, Hyderabad, Telangana, India
---------------------------------------------------------------------***----------------------------------------------------------------------
Abstract – Artificial Intelligence is concerned in developing Strong AI tends to think like a human and process like a
machines or systems that can perform human tasks with human to think more creatively and achieve the best
precision. These systems are highly efficient and are capable of possible outcome whereas a weak AI only tends to achieve
showing intelligent behavior. These systems learn from their result like a human by processing its database before making
experiences. AI is a rapidly developing area of computer a decision or a step towards the result.
science and Technology. AI has a huge scope in future,
currently many researches are going on in the field of Most of the Artificial Intelligence machines are based on
Robotics, Machine Learning, Neural Networks and Pattern expert systems. Expert systems are practically a knowledge
Recognition. Expert Systems are an important part of Artificial database which gives support and information needed to the
Intelligence. Expert systems help the AI in making decisions AI to make important decisions or to solve problems. AI is
based on intelligence. AI has the ability to reshape the future further divided into categories like machine learning, rnn,
of innovations and it can also impact the Research and patter recognition, etc. object identification is also an
Development sector in a positive way. In this paper I am going important part of AI, in this the system is a fed a lot of
to discuss about Artificial intelligence and its applications in information and objects to memorize which is further used
industries along with a brief on expert systems. AI has already to recognize when it is prompted this is used in facial
started making its way into the automobile industry with self- scanning security systems it is also used in a popular system
driving cars, navigation and many more by using pattern called ‘Face Id’ by Apple. AI is supposed to learn from day to
recognition. day life to make itself better there are two types of AI based
on learning
Key Words: Artificial intelligence, Applications, Expert
systems, Pattern Recognition.  Supervised learning

1. INTRODUCTION  Un supervised learning

AI also known as Artificial intelligence is a branch of In supervised learning the system is provided with a
Computer Science it mainly focuses on creating machines scenario and a result which leads to outcome. This relies on
that have the ability to think and solve problems. AI also has database.
intelligence like humans which helps to think like them in
situations to get a possible and likely outcome in a particular In unsupervised learning the system is given a scenario
scenario. The fascination of human brain has led to a without the result so the machine thinks and gives out a
question Can machines think? This started the development possible outcome. This relies on intelligence of the system in
of AI. John McCarthy the father of Artificial Intelligence said order to train it.

“Every aspect of learning or any other feature of intelligence During the initial days of development of AI there were some
can in principle be so precisely described a machine can be important aspects of it they are as follows:
made to stimulate it”. This is the main principle and idea
behind Artificial Intelligence. AI is also a type of computer  Simulating complex functions of human brain.
program which relies on a knowledge database it learns
more from its tasks and it is self-developing. AI is connected  Natural language processing.
through neural networks it can be also programmed to
 Self development.
create music, create movies. Main goal of AI these days is to
create a system which contains creativity, intelligence and  Knowledge.
randomness which a very difficult step and a challenge for
the future. Industries have moved from just developing  Randomness and creativity.
intelligent machines which provide solutions to creating
machines which can think randomly and logically. AI can be While most of them have been successfully achieved
categorized to two types: randomness and creativity are yet to be completed. Research
is ongoing in this area of development which will lead us to
 Strong AI an AI system much more capable than a human brain.
 Weak AI

© 2020, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 2195
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 08 | Aug 2020 www.irjet.net p-ISSN: 2395-0072

Fields where Artificial intelligence is used are: 1.2 Disadvantages of AI

 Health care There are some disadvantages of Artificial Intelligence they

are as follows:
 Automobile industry
 AI systems are very expensive to develop.
 Agriculture
 They are very expensive and highly complex to
 Smartphones, etc. maintain.

1.1 Advantages of AI  They require more data during machine learning to

achieve more accuracy.
As we know there are many advantages and
disadvantages of AI. Some of its advantages are as follows:  The information sources we humans have is limited
to a certain extent, so we cannot provide complete
 Artificial intelligence as any machine mainly information to the AI system hence they cannot
reduces errors by working in a systematic way replicate human level of intelligence.
and they won’t be affected by any hostile
environment which results in more precession.  Unemployment of humans.

 As AI is a machine it can increase productivity  AI cannot be creative.

by performing more work because it is not
subjected to tiredness or fatigue like humans.  Increase in costs.

 Ai machines can take over repetitive jobs which  The programmers who have knowledge in AI are
are monotonous in nature thereby expanding very few.
human creativity.
 Development process of AI systems is very slow
 Ai is also very useful in field of health care. when compared to other technologies.

 It also makes faster and accurate decisions  It is very hard a computer to achieve human like
intelligence in all areas.
 Ai can detect fraud easily and prevent any
losses.  Complex coding.

 It is also very interactive.  They lack creativity and randomness.

 Used in warfare to avoid human losses.  Mental health and mental capabilities of humans
become too dependent on AI if used always.
 Can adapt to any changes easily.
 AI systems lack wisdom.
 They have improved interfaces which are slightly
easy to handle.  Machines can be easily misused which is a threat to
humanity and human life.
 Converts information into knowledge easily.
AI having human level of intelligence is questionable:
 AI systems think logically. ethically can we provide our intelligence to a machine with
no emotions as intelligence to humans is considered a gift of
As they can think logically without emotions, partiality nature.
they can make accurate and precise decisions. AI systems
can be very useful in assessing people with difficulties of 1.3 Key statistics of AI
any sort. They are helpful in assessing surgeons perform
radiosurgery. They don’t need rest as they don’t get tired  Today 37% of all business organizations employ AI.
which increases productivity which humans can’t achieve
as it is impossible.  Voice assistants which use AI are around 3 billion.

 AI industry might earn a total of $118 billion every

year by 2025.

© 2020, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 2196
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 08 | Aug 2020 www.irjet.net p-ISSN: 2395-0072

 Amazon’s voice assistant alexa is powered by AI Self-aware AI it has human like intelligence and self-
technology which has around 66,000 skills in its awareness. This type of AI is also currently in development
inventory. stage.

 Self driven cars are safer than regular cars in the US. 1.4 Goals of AI:

Artificial intelligence has also impacted the job marked in a There are two types of goals of AI they are:
huge way. There are around 133 million jobs which will be
created by presence of Artificial Intelligence.  Scientific goal: used to understand about the
mechanism of human intelligence.

 Engineering goal: building a tool or machine with

human intelligence to solve real world problems.

Examples for engineering goals of AI:

 Reproduce human intelligence in machines or

systems.

 Solving knowledge intensive tasks which require lot

of information.

 Intelligent connection of perception and action.

 Speech and vision recognizing systems.

Chart: Jobs impacted by AI.  Game playing systems.
This shows the jobs that will be impacted by AI in  End of human mankind by Stephen Hawking.
percentage. AI will increase jobs in some fields whereas it
might take away some jobs in certain fields.  Life 3.0 by Max Tegmark.

1.3 Types of AI  Silicon consciousness by Michio Kaku.

There are 4 types of Artificial intelligence they are: 1.5 Applications of AI
 Reactive AI AI is being used in many fields such as:

 Limited memory AI  Games – AI has been used in many games which

involve a lot of thinking and strategies. Games such
 Theory of mind AI as chess, poker, etc. AI system can think many
moves ahead where it can predict more than one
 Self-aware AI possible outcome.
Reactive AI has no memory but responds to different  Financial sector – Many banking organizations use
stimuli. They do not have the capability to learn. AI to organize many account information. Used to
invest in stocks and detect any changes.
Limited Memory AI they contain memory but only to some
extent. It can learn from previous experiences and improve
 Music – sound processing, composition, etc.
its responses.
 Automobile – automatic transmissions where cars
Theory of Mind AI this is a work in progress AI which
can be automatically driven without a driver or any
researchers are currently developing like humans this
gear changes.
contains emotions, wisdom, logic, etc. it understands the
needs of user.  Hospitals – AI is used in hospital by staff and
surgeons in organizing bed schedules, detecting
tumors, performing surgeries, organizing beds, etc.

© 2020, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 2197
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 08 | Aug 2020 www.irjet.net p-ISSN: 2395-0072

 Expert systems – Where certain machines are used  Misuse of this technology is a big threat.
by combing practical knowledge and intelligence
these machines provide advises to the users.  99% of jobs might be eliminated in the future based
on AI.
 Natural language processing – It is used in many
voice assistants which makes it possible for a  It might create better environment between
human to interact with a computer or a machine in organizations and customers.
normal language.
 There is a bigger question in AI development; do we
 Intelligent robots – Where robots are programmed have enough knowledge resources for it to replace
to perform certain tasks given by humans. They human expertise and intelligence?
have high intelligence and knowledge and they can
also learn from their own mistakes which makes  AI might not have enough creativity to think like a
adapting to a new environment easier. They contain human brain it will be more goal oriented so it can
many sensors which detect heat, light, any bumps easily be used in destructive mode.
on the surface, etc. which are necessary for
performing tasks and its survival in any kind of  If placed in wrong hands AI can be very dangerous
environment it is placed in. to humanity in the future considering it’s abilities
and knowledge base.
1.6 Real life uses of AI
 Work load on humans will be minimized as AI can
 OTT platforms: AI is used to recommend the take over some jobs from us.
movies or any shows that a person might watch
based on their previous watched shows. As we see there are many beneficial impacts in development
of AI and similarly there are many downsides to it. We
 Drones: AI is applied in drones where they fly on should be very careful in this process of transition and in the
their own by using sensors and cameras and they future.
recognize paths or faces and have the ability to
follow it on its own without giving any controls. 2. EXPERT SYSTEMS

 Self driving cars: In the process of automation AI Expert System is an application which is part of AI. It
has been implemented in cars to drive on its own practically consists of a knowledge database which helps in
using many sensors it uses cameras to recognize the solving complex problems and make complex decisions.
road and obstacles to drive itself. An example of this Expert System mimics human decision making ability in
technology is tesla cars. Artificial Intelligence systems. It consists of high knowledge
and expertise. Expert systems mainly represent if-then rules
 Natural language processing: used in voice rather than giving a procedure code in AI. It is one of the best
assistants where we can interact with a computer or researches in AI field. It has been introduced by researchers
a device in our own language, also used in speech at Stanford University.
recognition.
2.1 Characteristics of Expert Systems
 Robots: Theses AI powered machines are found to
 Expertise: They contain more knowledge so
be a helping hand to humans in healthcare sector in
decisions will be accurate and correct and the cost
many areas.
of accessing information is reduced.
 Google: Initially it was not an AI but due to its
 Symbolic reasoning: It is easier to validate
continuous development and updates it has become
information as Expert system uses symbols for
the world’s largest AI.
representation.
1.7 Impact of AI
 Reliable: They are very reliable to access any kind
There is a bright and a dark future in this sector which is of information.
developing slowly but steadily.
 High performance systems.
 Transition to AI will be extremely hard and painful.
 Easily understandable.
 It cannot completely takeover human as there are
some jobs which only human can do.

© 2020, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 2198
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 08 | Aug 2020 www.irjet.net p-ISSN: 2395-0072

2.2 Capabilities of Expert System

 Accurate advises

 Arriving at a certain solution

 Understanding

 Predicts different results

 Suggestions in alternate methods to solve a

problem.

 Assisting humans in different situations

 Justifying the result Fig-1: Components of Expert Systems

2.3 Components of an Expert System 2.4 Limitations of Expert Systems
There are basically three main components of an Expert Like every technology ES also have certain limitations they
System: are:
 Knowledge base  Development costs are high.

 Inference engine  Difficult to maintain.

 User interface  Less people with expertise in ES.
In knowledge base we usually have information, data, facts  Requires more knowledge.
which are widely accepted around the globe by knowledge
scholars and engineers. Knowledge base is required to 2.5 Benefits of Expert Systems
exhibit decisions and solutions. We can also say that the
success of an ES is mainly based on the kind of knowledge it  Production cost is less.
acquires, the more and precise knowledge the better the ES
 Easily available.
It also consists of past experiences from which it can
improve and develop  Reliable.

Knowledge in an ES is mostly represented by IF-THEN-ELSE  Reduce effort for individuals.

rules which are used to formalize the knowledge.
 High speed.
Inference engine makes use of certain procedures and rules
based on the knowledge an ES has acquired to arrive at a  Large source of information.
correct and flawless solution for a given problem
 Minimize training costs in industries.
User interface provides the interaction between user and
the ES itself by which they can interact with the system. It is  Works without emotions and fatigue.
also helpful in explaining the steps followed by the system to
arrive at a solution. It is similar to natural language 2.6 Different types of Expert Systems
processing. For a user to access an ES they need not be well
equipped with knowledge in expert system. This makes it There are mainly five types of Expert Systems they are:
easier to understand the solutions and access them easier
 Rule based expert system.
 Fuzzy expert system.
 Frame based expert .system
 Neural expert system.
 Neuro-fuzzy expert system.

© 2020, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 2199
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 07 Issue: 08 | Aug 2020 www.irjet.net p-ISSN: 2395-0072

2.7 Applications of Expert Systems really with full of surprises and new ideas this is an area of
technology which is not completely yet developed but just
 Diagnosis systems to detect cause of a disease from the process of thought in machines is a huge leap forward in
given data, conduction of medical operations on technology.
humans.
REFERENCES
 Finding out faults in any kind of device like
computers, vehicles, etc. 1) https://www.tutorialspoint.com/artificial_intellige
nce/artificial_intelligence_overview.htm
 Planning and scheduling. (“Introduction and Goals”)

 Knowledge publishing. 2) https://futuremonger.com/what-is-the-

ultimategoal-of-artificial-intelligence-
 Loan analysis. e9cd4d07e172 ("Goals of AI")

 Virus detection. 3) https://www.researchgate.net/post/advantages_an

d_disadvantages Artificial_Intelligence_AI
 Repair and maintenance. (“advantages and disadvantages of AI”)

 Marketing. 4) http://www.umsl.edu/~joshik/msis480/chapt11.h
tm ("Expert systems")
There are many applications of Expert systems and Artificial
Intelligence. 5) https://www.tutorialspoint.com/artificial_intellige
nce/artificial_intelligence_expert_systems.htm
2.8 Examples of Expert Systems ("Expert systems & applications")
 MYCIN: it is an ES based on backward chaining it 6) https://www.tutorialspoint.com/artificial_intellige
can identify certain bacteria which cause acute nce/artificial_intelligence_expert_systems.htm
infections and it is also used to prescribe drug (“Expert systems & applications”)
dosage based upon patient’s weight.
7) http://cdn.differencebetween.net/wp_real_timeapp
 DENDRAL: used to predict molecular structure to lications (“Applications in real life”)
perform chemical analysis.
8) https://www.forbes.com/sites/cognitiveworld/201
 PXDES: it predicts the type and degree of lung 9/06/19/7-types-of-artificial-
cancer in patients. intelligence/#1e700232233e (“types of AI”)

 CaDet: it can identify cancer at early stages in 9) https://pdfs.semanticscholar.org/973a/c1c845241

patients. 5fa16dea80b694574ceed0d4364.pdf (“types of ES”)

3. CONCLUSION 10) https://dataprot.net/statistics/ai-statistics/ (“key

statistics of AI”)
Now, the time to think on the future of artificial intelligence
and expert systems think that as to go with normal 11) https://en.wikibooks.org/wiki/The_Computer_Rev
programming or adapt to artificial intelligence. AI has olution/Artificial_Intelligence/Expert_Systems
currently been a useful tool in the software industry. AI in its (“benefits of Expert systems”)
current short existence has led to many improvements in
technology and has improved quality of life for humans. It BIOGRAPHIES
has provided an impressive range of applications. At the
same it has shown us the challenges that we may face in the Nagalla Satya Abhisht currently
future due to its development. AI has also provided many pursuing III B.Tech at GITAM
new areas of study to explore and develop for future University, Hyderabad. My
generations. It has provided a new way of human thinking. research areas are data analytics,
The overall motivation behind this paper is to understand data mining, and propulsion in
the developing technology which is necessary in order to aerospace.
cater the needs of growing population. But this concept
requires a paradigm shift in human thinking which changes
the way we think about modernization based on needs more
than modifying existing techniques. The subject of AI is

© 2020, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 2200