Natural materials display a structural hierarchy that spans from the nano-scale to the macro-scale, with some of their unique properties partially driven by nanostructural anisotropy. In this research, we build upon our recently established technique of mechanically-directed assembly, examining its capacity to fabricate multi-functional, porous biopolymers with engineered anisotropy at the nano-scale in three dimensions. To do this, we employ 3D-printing technology to create complex molds made of silicone. Into these molds, we infiltrate cellulose and alginate prepolymer solutions, which are then allowed to solidify into a gel form. By subjecting these polymers to controlled exposure to a polar solvent, we can apply both precision mechanical strain and physical crosslinking to the polymers. This precisely applied strain leads to the programmability of anisotropy within the resultant material. In the final step of the process, we implement a critical-point drying technique, resulting in the formation of anisotropic, nanofibrillar aerogels. These aerogels display vibrant patterns of birefringence that are a direct result of the alignment of nanofibers within the material. Our method appears to be universally applicable across physically-crosslinked polymers. Additionally, it can be utilized to permanently align co-infiltrated functional nanowires within the polymers. In summary, our technique of mechanically-directed assembly allows for the simple, cost-effective production of hierarchical materials with meticulously engineered properties. These properties are structure-dependent and span the chemical and polymeric realm to include electronic, magnetic, and mechanical characteristics.

In the study, we analyzed variability of three univariate and multivariate drought indicators and yields of five of the largest rain-fed crops in Australia including wheat, broad beans, canola, lupins and barley. Using multivariate copulas, this study relates changes in climate variability to changes in crops production during 1980-2012. In the analysis period, the five chosen crops indicate a modest association with the selected drought indicators: Standardized Precipitation Index (SPI); Standardized Soil Moisture Index (SSI), and Multivariate Standardized Drought Index (MSDI). The latter combines precipitation and soil moisture and provides a measure of agro-meteorological drought. A model is developed to describe the relationship between drought and crop production using copulas. The model offers the likelihood of crop yield given an observed or predicted SPI, SSI or MSDI.

The rapid advancement of powerful Large Language Models (LLMs), such as ChatGPT and Llama, has revolutionized the world by bringing new creative possibilities and enhancing productivity. However, these advancements also pose significant challenges and risks, including the potential for misuse in the form of fake news, academic dishonesty, intellectual property infringements, and privacy leaks. In response to these concerns, this thesis explores approaches to promoting the responsible use of LLMs from both theoretical and empirical perspectives.

Three key approaches are presented: (1) Detecting AI-generated Text via Watermarking: We propose a robust and high-quality watermarking method called Unigram-Watermark and introduce a rigorous theoretical framework to quantify the effectiveness and robustness of LLM watermarks. Furthermore, we propose PF-Watermark, which achieves the best balance of high detection accuracy and low perplexity. (2) Protecting the Intellectual Property of LLMs: We safeguard the intellectual property of LLMs through novel watermarking techniques designed to prevent model-stealing attacks in both text classification and text generation tasks. (3) Privacy-Preserving LLMs: We employ Confidential Redacted Training (CRT) to train and fine-tune language generation models while protecting sensitive information. In summary, we propose a suite of algorithms and solutions to address LLMs' trending safety, security, and privacy concerns. We hope our studies provide valuable insights for researchers to explore exciting future research solutions that promote responsible AI development and deployment.

Background

Results

Conclusions