Upcoming Doctoral Dissertations

"Graph learning is a rapidly growing field in machine learning. Despite its success, there are several concerns about the trustworthiness of graph learning models. In my research, I focus on two important issues in trustworthy graph learning: fairness and privacy. Fairness aims to mitigate bias introduced or amplified during the learning process. Privacy requires that the privacy of the data and model parameters, which are regarded as confidential information belonging to their owners, should be protected.." Read more...

"Learning homomorphisms with noise (LHN) is a group-theoretic learning problem generalizing quantum safe computational assumptions like learning parity with noise (LPN) and well-established learning with errors (LWE). The LHN problem associated with Burnside groups of exponent three is referred to as learning Burnside homomorphisms with noise (Bn-LHN). In a nutshell, the Bn-LHN problem focuses on recovering a secret homomorphism between Burnside groups, given polynomially many noisy samples. Previous work assessed important combinatorial properties and basic cryptographic applications of the Bn-LHN problem, but did not address efficient constructions of a fundamental cryptographic primitive known as pseudorandom function (PRF). " Read more...

"Patterning surfaces with desirable structures and multifunctionalities and at nano- and micro-length scales is essential for engineering functional materials. This research explores advanced applications of electron-beam lithography (EBL), specifically focusing on two distinct aspects: the design of surface-tethered shape morphing microgels and surface pattering with orthogonal functionalities. EBL offers high-resolution, mask free patterning at nano-length scales by exposing polymeric precursors to energetic radiation. Firstly, diverse incident electron energies are employed for the fabrication of shape-morphing microgels from homopolymer films of poly (acrylic acid) (PAA). " Read more...

"Today's energy and power systems generate many societal benefits, such as reliable and affordable electricity, jobs, and economic growth, but they also produce harmful emissions and waste. Promoting a deeper deployment of renewable energy sources such as solar and wind power is one approach to reducing the environmental impact of the electricity grid, although it raises concerns about energy costs and grid reliability." Read more...

"Experiential knowledge is the domain of the technical expert. Derived from experience, experiential or tacit knowledge differs from a priori knowledge learned through reading or specification. Difficult to ascertain or describe, expert knowledge is the domain of knowledge management systems and the subject of this research. The systems engineering domain has been held accountable for a lack of elegance in system design. Resolving this criticism requires reducing the cognitive load of the performing architect, often in cases with little historical precedence. " Read more...

"In the past decade, Artificial Intelligence (AI) has experienced a notable renaissance. AI’s capabilities in parsing complex systems have been particularly evidenced in the advances within Natural Language Processing (NLP). The introduction of the Word2vec algorithm in 2013 offered a means to translate text into numerical vectors and catalyzed the application of convolutional and recurrent neural networks to text analysis. In 2017, Transformers revolutionized the NLP field and started a new chapter in AI with large language models." Read more...

"The study of financial contagion is paramount to safeguarding the stability of the global financial system. To understand the mechanisms and channels which propagate financial contagion, existing studies propose 'inverse demand functions' to mathematically model the relationship between trading activities and resultant prices, measuring the impact of asset sales on their prices in financial contagion. We have employed a network model based approach to study the inverse demand function within the context of financial contagions." Read more...

"Generative language models (GLMs) represent a new class of large language models (LLMs) that specialize in generating text. As such, these models present new opportunities in finance due to their ability to generalize learned knowledge through conversational dialog without explicit training. While this ability expands the range of applications of LLMs for finance, text-based outputs are ill-suited for traditional cross-sectional portfolio construction methods, which rely on the ability to rank expected return forecasts for portfolio selection." Read more...