Upcoming Doctoral Dissertations

"Modern satellite remote sensing plays a crucial role in providing data on various water, atmosphere, and land surface conditions. This research introduces improvements to remote sensing methods through a new method for quantifying measurement uncertainties in atmospheric correction algorithms of an existing tool for retrieval of aerosol and marine parameters from ocean color data (OC-SMART); an exploration of the impact of hyperspectral versus multispectral data channels on snow parameter retrieval algorithms; and applications of OC-SMART to Arctic water inherent optical property retrievals. Chapter 1 contains a background on remote sensing of environments; chapter 2 discusses critical tools used in this research; chapter 3 describes how to quantify uncertainties in OC-SMART using Bayesian inversion; chapter 4 explores the impact of hyperspectral information on retrievals of snow grain size and impurity concentration; chapter 5 discusses the application of OC-SMART to Arctic water inherent optical property retrievals; and chapter 6 summarizes the research and provides closing remarks." Read more...

"The fields of tissue engineering and regenerative medicine have made astounding progress in recent years, evidenced by cutting-edge 4D printing technologies, precise gene editing tools, and sustained longterm functionality of engineered tissue grafts. Despite these amazing feats, clinical success of tissueengineered constructs is still limited to relatively simple tissues. Volumetric tissues (typically those larger than a few millimeters) suffer from poor oxygen supply due to limited diffusion. Large, complex tissues require a vascular network to supply the growing cells with nutrients needed to meet the metabolic demands of prolonged viability and tissue regeneration." Read more...

"Structural health monitoring is vital for the early detection of structural damage, enabling timely and cost-effective rehabilitation. Traditional structural health monitoring approaches suffer from limitations such as low accuracy, reliance on manual intervention, inefficient data analysis, and susceptibility to human error. This dissertation presents the development of novel machine learning based sensing techniques to monitor both macroscopic and microscopic structural behaviors, thereby enhancing structural health monitoring capabilities and offering a pathway to intelligent infrastructure systems." Read more...

"Polymer entanglements play a critical role in influencing the viscoelastic properties of polymer nanocomposites. Factors such as crosslinking, chain adsorption onto nanoparticles, and chain architectures (such as linear and loop grafting) dictate the degree of entanglement within these composite materials. This thesis aims to explore the influence of entanglements on the viscoelastic properties of three systems with different topological constraints. In the first system, we synthesized entangled and unentangled composite networks to elucidate the impact of both entanglements and particle loadings on the reinforcement. The study demonstrated that the elastic moduli of crosslinked composite with entangled chains remained unaffected with particle loading, whereas both the elastic moduli and crosslinking degree were enhanced with increasing particle loading in unentangled crosslinked composite. " Read more...

"Machine learning has been a powerful tool in the modern world. In the past decades, due to the explosion of unverified data sources, and the increasing interaction between human and computer, it is of concern whether machine learning algorithms are robust to data corruption or even adversarial attacks. On the other hand, in many real-world scenarios, it is hardly the case that we can gather enough data for training a good model, making data-efficiency another important aspect of algorithmic designs.." Read more...

"The main objective of this work is to address the evaluation and optimal control of risk of distributed systems. Our motivation stems from problems arising in robotics, where a team of robots cooperates to complete a common task in an unknown environment. In these types of problems, the risk may be associated with systemic features pertaining to the whole system as well as with features related to the local operation of individual agents. The goal is to devise methods for evaluating and optimizing the total risk by integrating both systemic and local features while allowing independent operation of agents within the system under limited information exchange." Read more...

"Quantum dynamics is the study of the time evolution of quantum observables with respect to a system, which in the context of this work, is a few-level atomic system interacting with a set of external EM fields. We study light-matter interaction in a number of interesting systems, with special focus on the occurrence of multiphoton processes on ultrafast timescales and how particular states of light can lead to desired dynamics of the atomic system and propagating light. Such an endeavor is the objective of quantum control, where we study the processes that underlie outcomes of interest for both the atomic system and light field, and seek to determine protocols that realize those outcomes. Such outcomes include selective population transfer in a manifold of atomic states or controlling the steady state response function of the atomic system to realize optical phenomena such as induced transparency and enhanced emission, all of which we study in this work." Read more...

"Security analytics (a.k.a. log analysis) is a prominent information-based approach for defending enterprises against cyberattacks. Organizations often employ third-party Security Information and Event Management (SIEM) systems to collect, analyze, and manage log data that originate from various Security Analytics Sources (SAS) to identify and mitigate threats in real-time, investigate incidents, troubleshoot errors, and improve their security posture. However, the efficiency and effectiveness of an SIEM depend heavily on the quality and quantity of the provided logs, and adversaries often tamper with logs after an intrusion to cover their attack traces. Thus, protecting the confidentiality, integrity, and availability of logs, whether at rest, transit, or execution, becomes vital for the efficacy of any security analytics solution." Read more...

"Lithium metal has gained extensive attention to overcome the theoretical limitations of intercalation-type anodes. Unlike commercial graphite anode, metallic lithium is well-known for its high energy density and low reduction potential. The anode-free configuration with zero-excess lithium is progressively suggested as a particular type of lithium metal battery for safe and facile manufacturing purposes. However, due to dendritic lithium formation, realizing poor cycling stability and Coulombic Efficiency remains a continuing challenge. " Read more...

"Mixtures are ubiquitous in chemistry, ranging from simple combinations of two or more substances to complex formulations encountered in various natural and synthetic contexts. Mixture analysis involves the identification and quantification of individual components within a mixture, elucidating their concentrations, interactions, and structural characteristics. Modern instrumental techniques, including chromatography, spectroscopy, and mass spectrometry (MS), have been developed to address the diverse challenges posed by mixture analysis. In this dissertation work, the procedures of ultra-performance liquid chromatography (UPLC) MS and ion mobility (IM) MS used to analyze mixtures were investigated." Read more...

"Metal nanoparticles (MNPs) as a major nanomaterials category possess completely different properties compared to bulk materials and ion forms, which inherit and expand the possibilities from their original elemental counterparts. MNPs exhibit excellent optical, electrical, and catalytic properties, benefiting from mature synthetic and conjugation methods, as well as their high surface area-to-volume ratios. They also demonstrate unique magnetic, thermal, and quantum confinement effects. This diverse range of tunable properties allows metal nanoparticles to be intensively utilized in diagnosis, catalysis, drug delivery, biomedical applications, etc. Modification of particle surfaces and coordination with other nanomaterials can further optimize their capabilities and advantageous characteristics, enabling application-driven design and engineering." Read more...

"Lithium Niobate is revitalized in the field of photonics due to its outstanding optical properties, including a broad transparency window, low propagation loss, and high nonlinearity on a thin film platform. These properties position Lithium Niobate as a competitive candidate to become ‘the Silicon of photonics’. To explore the applications of thin film Lithium Niobate, in this thesis, I demonstrate the workings of a High-Q resonator and single-photon nonlinearity on thin film Lithium Niobate platform. A high-Q resonance of 4×106for transverse-electric mode is reached with specialized pattern design and fabrication process. To fully use the high Q resonance, periodical poled Lithium Niobate is introduced for its high nonlinear coefficient 𝜒2." 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...

"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...

"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...

"In today’s rapidly evolving world, intelligent systems are taking on a crucial role in assisting human across various fields, spanning healthcare to transportation. Within this context, the level of user reliance becomes a crucial factor determining how effectively intelligent systems can support human decision-makers innumerous domains. How much a user should rely on an intelligent system to maximize the benefits is an open question. In this research we try to study the dynamics of human – AI interaction, aiming to investigate the optimal level of reliance that lead to effective collaboration. In this research we utilized a multidimensional approach that combines computational modelling and human subject experimentation to unfold the research questions. Through the lens of computational simulation, we studied the optimal reliance level that is needed to gain the maximum performance in human-decision support system teaming. By analyzing various characteristics, our research unearths how user should appropriately rely on intelligent systems to maximize collaboration performance. This novel perspective helps system designer to tailor AI systems effectively" Read more...

"This dissertation seeks to clarify how community biology labs as social enterprises can play a role in optimizing higher-order systems in line with their social goals. Utilizing these labs to reroute the exogenous insulin ecosystem in the US seems a promising application of citizen science. However, there are a number of barriers to their participation in healthcare and life science research. With their inherently individualized organizational structure, community labs face a range of common and unique obstacles. In this dissertation, I use an approach grounded in systems science to study social distributed systems and propose solutions to these challenges." 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...

"Extensive research has demonstrated that qualitative information in managerial disclosures holds significant value for understanding financial markets and events. Traditional textual analysis methods often rely on predefined dictionaries to measure qualitative information without adequately considering its contextual nuances. To address this limitation, my dissertation employs contextualized methods based on deep learning to quantify underexplored qualitative information or hidden cues from managerial disclosures. Specifically, it focuses on: (1) granular changes of 10-K reports and (2) analysts’ skeptical attitudes in earnings calls. Empirical experiments show that these hidden cues provide predictive capabilities for financial fraud and market activities." Read more...

"Agent-based modeling (ABM) is a simulation approach that is increasingly used in studying household finance and banking systems. ABM allows for creating a virtual environment where agents (individuals or institutions) interact with one another and make decisions based on a set of rules and objectives. In this dissertation, we discuss three essays about ABMs in both household finance and banking systems. The first essay focuses on ABM in household finance. We model households as economic agents making financial decisions through their life cycle. More specifically, in our life-cycle model, individuals facing uninsured labor income risks choose whether to participate in the stock market and decide on home ownership in an environment with a social safety net and retirement savings system" Read more...

"The three essays in my dissertation focus on corporate strategies: the impact of communication technology on corporate diversification, the effects of CEO overconfidence on equity issuance, and the role of management and media sentiments in mergers and acquisitions. The first essay, “Analyzing the Corporate Diversification Strategy in the Era of Communication Revolution”, focuses on the impact that advancements in communication technology have on the corporate structure. It reveals a trend among 10,640 companies (1998-2019) towards reduced vertical integration and diversification as communication technology improves, allowing companies to focus more on core competencies and outsource non-core functions. The research also explores how these technological advancements impact corporate strategies related to R&D, intangible assets, and the roles of young CEOs, finding that technology advancement softens the influence of these three factors on a firm’s structure. Upon conducting a textual analysis of the MDA sections within firms’ 10K filings, we also observe that as communication technology improves, companies show less explicit interest in technology and outsourcing." Read more...

"The three essays presented in my dissertation proposal explore some key issues in asset pricing: horizon pricing, stock momentum, and factor momentum. In the first essay, “Arbitrage Asymmetry, Mispricing Gap, and Momentum”, I show that the mispricing score gap (Mgap) between overpriced winners and underpriced losers effectively measures the level of persistence in the trading behaviors of arbitrageurs and sentiment investors simultaneously and is a strong predictor of stock momentum. This predictability is statistically and economically significant both in-sample and out-of-sample. The in-sample predictability remains even when accounting for various state-of-the-art common risk factors and existing predictors. A one standard deviation increase in Mgap boosts next month’s momentum returns by 1.10%, which is over 90% of the historical 1.16% return. This predictability seems to stem from market efficiency determined by the dynamics of both sentiment investors and arbitrageurs." Read more...

"The three essays presented in my dissertation aim to mitigate the impact of estimation error in portfolio optimization, specifically for the mean-variance strategy proposed by Markowitz. The first essay, “Partial Index Tracking Enhanced Mean-Variance Portfolio” proposes a partial index-tracking strategy that better approaches the true mean-variance portfolio compared with the sample approach. We replace the portfolio variance in the mean-variance optimization with the linear combination of portfolio variance and its tracking error. In order to find the optimal tuning parameter, we take advantage of the Expected Out-ofsample Utility (EOOSU) framework proposed by Kan & Zhou and minimize the mean square error with respect to the true mean-variance rule. The empirical results show that the proposed rule can provide similar performance as the other widely used strategies in literature. In the meantime, it delivers relatively smaller turnover and tracking error compared to other candidates.." Read more...