<![CDATA[Ph.D. Dissertation Defense - Alexander Benvenuti]]>28475Title: Differential Privacy for Symbolic Systems
Committee:
Dr. Matthew Hale, ECE, Chair, Advisor
Dr. Samuel Coogan, ECE
Dr. Saman Zonouz, ECE
Dr. Miriam Kennedy, AFRL
Dr. Sarah Li, AE
]]> Daniela Staiculescu117852784352026-07-28 22:40:3517852784952026-07-28 22:41:3500eventThe purpose of this dissertation is to develop a suite of privatization algorithms for symbolic systems, specifically systems modeled as Markov chains, hidden Markov models, and Markov decision processes, to protect the sensitive user data used to generate these models. We use differential privacy as our framework to provide privacy protections to user data. To enforce differential privacy for symbolic systems, we develop algorithms for privatizing reward functions, model constraints, and transition dynamics. Additionally, we develop a framework for privatizing the trajectories generated by these models, a filtering framework for these trajectories. For each framework, we provide users with tools to calibrate the strength of privacy and to analyze the accuracy of the privatized data. Future work will apply the algorithms in this dissertation to domain-specific problems, and extend these algorithms to provide privacy for continuous-time symbolic systems.]]>2026-08-10T15:00:00-04:002026-08-10T17:00:00-04:002026-08-10T17:00:00-04:002026-08-10 19:00:002026-08-10 21:00:002026-08-10 21:00:002026-08-10T15:00:00-04:002026-08-10T17:00:00-04:00America/New_YorkAmerica/New_Yorkdatetime2026-08-10 03:00:002026-08-10 05:00:00America/New_YorkAmerica/New_Yorkdatetime<![CDATA[]]><![CDATA[Ph.D. Dissertation Defense - Ken Li]]>28475Title: High-Resolution Digital-Alike ADC Architectures and Design Automation
Committee:
Dr. Shaolan Li, ECE, Chair, Advisor
Dr. Jane Gu, ECE
Dr. Shimeng Yu, ECE
Dr. Visvesh Sathe,ECE
Dr. Levent Degertekin, ME
]]> Daniela Staiculescu117852778642026-07-28 22:31:0417852779352026-07-28 22:32:1500eventThis dissertation presents high-resolution digital-alike analog-to-digital converter (ADC) architectures and design automation techniques for improving energy efficiency, scalability, and design productivity in advanced CMOS technologies. First, an automated design methodology is developed for first- and second-order VCO-based delta-sigma ADCs. The proposed framework integrates architecture-level exploration, circuit sizing, and simulation-based optimization to reduce the manual effort and design time required for analog and mixed-signal circuits. Second, a third-order continuous-time delta-sigma modulator using a phase-time two-step quantizer is presented. By combining VCO-based and time-domain quantization techniques, the proposed architecture achieves high-resolution conversion with low power consumption. Finally, a high-resolution pipeline-SAR ADC is developed using a driver-relaxed input interface, capacitor mismatch shaping, and a power-efficient closed-loop dynamic residue amplifier. These techniques improve input drivability, robustness, and conversion efficiency without requiring intensive calibration. Together, these works demonstrate how digital-alike circuit techniques, architectural innovation, and design automation can enable scalable and energy-efficient high-resolution ADCs.]]>2026-08-12T09:30:00-04:002026-08-12T11:30:00-04:002026-08-12T11:30:00-04:002026-08-12 13:30:002026-08-12 15:30:002026-08-12 15:30:002026-08-12T09:30:00-04:002026-08-12T11:30:00-04:00America/New_YorkAmerica/New_Yorkdatetime2026-08-12 09:30:002026-08-12 11:30:00America/New_YorkAmerica/New_Yorkdatetime<![CDATA[]]><![CDATA[Microsoft Teams Link ]]><![CDATA[Ph.D. Dissertation Defense - Brett Ringel]]>28475Title: Design and Characterization of Photonic and RF Integrated Circuits and Devices for Harsh-Environment Space Missions
Committee:
Dr. John Cressler, ECE, Chair, Advisor
Dr. Nima Ghalichechian, ECE
Dr. Stephen Ralph, ECE
Dr. William Hunt, ECE
Dr. James Wray, EAS
]]> Daniela Staiculescu117852776672026-07-28 22:27:4717852777692026-07-28 22:29:2900eventThis thesis explores the work performed towards characterizing, understanding, and mitigating observed degradation in radio frequency and photonic integrated circuits and systems for use in high-radiation space environments. Specifically, work towards creating and characterizing integrated radiation-hardened by design complementary-metal-oxide-semiconductor transistors, and comparisons of electrical versus optical radiation-induced degradation in vertical-cavity photo-detectors, are explored. Additionally, comparisons between the radiation responses of different photonic interconnects (silicon vs. silicon-nitride), and potential mitigation techniques for the observed damage, are shown. Finally, the vulnerability of small-form resonator-based electro-optic modulators to transient radiation effects, and the implications of this increased sensitivity on photonic data links, is confirmed contrary to the perceived radiation transient hardness of photonic components.]]>2026-08-11T10:30:00-04:002026-08-11T12:30:00-04:002026-08-11T12:30:00-04:002026-08-11 14:30:002026-08-11 16:30:002026-08-11 16:30:002026-08-11T10:30:00-04:002026-08-11T12:30:00-04:00America/New_YorkAmerica/New_Yorkdatetime2026-08-11 10:30:002026-08-11 12:30:00America/New_YorkAmerica/New_Yorkdatetime<![CDATA[]]><![CDATA[Microsoft Teams Link ]]><![CDATA[Ph.D. Dissertation Defense - Ranjani Narayanan]]>28475Title: Investigating a Human-Centered Approach Towards Supporting Shared Mental Models in Hierarchical Human-Agent Teams for Decision Making
Committee:
Dr. Karen Feigh, AE, Chair, Advisor
Dr. Samuel Coogan, ECE, Co-Advisor
Dr. Sonia Chernova, CoC
Dr. Maegan Tucker, ECE
Dr. Zahra Ashktorab, Microsoft
Dr. Nancy Cooke, Arizona State
]]> Daniela Staiculescu117852774632026-07-28 22:24:2317852775522026-07-28 22:25:5200eventAdvances in AI have transformed decision-support systems from simple tools for humans into collaborative teammates. As humans increasingly supervise multiple heterogeneous agents, effective collaboration depends not only on AI capabilities but also on human ability to understand them. While prior research has focused on improving AI accuracy, transparency, and explainability, comparatively little is known about how humans develop cognitive representations of multiple AI teammates and use them to supervise complex teams. This dissertation investigates how humans develop and apply mental models of AI, i.e., Team Models, in hierarchical Human-Agent Teams (HATs). Through human-subject studies for human-agent hierarchical triads, this dissertation examines whether Team Models improve supervisory decision making, the factors that influence their development, and how different forms of prior information about AI shape Team Model formation and joint outcomes. The findings demonstrate that Team Models improve supervisory coordination, reduce workload, and enhance task efficiency, particularly when AI teammates provide conflicting recommendations. However, Team Model development is constrained by inter-agent dependencies, users' limited ability to identify AI failure modes, and the abstraction level of information presented by the system. The research further shows that prior information about teammate characteristics only selectively improves Team Model development, with its effectiveness depending on team structure. Moreover, increased knowledge about AI teammates does not necessarily produce calibrated reliance or improved collaboration, as users often rely on environmental feedback and simple heuristics rather than reasoning about teammate capabilities. Overall, this dissertation extends shared mental model theory hierarchical HATs and provides human-centered design guidelines that support better supervision and collaborative decision making.]]>2026-08-11T15:00:00-04:002026-08-11T17:00:00-04:002026-08-11T17:00:00-04:002026-08-11 19:00:002026-08-11 21:00:002026-08-11 21:00:002026-08-11T15:00:00-04:002026-08-11T17:00:00-04:00America/New_YorkAmerica/New_Yorkdatetime2026-08-11 03:00:002026-08-11 05:00:00America/New_YorkAmerica/New_Yorkdatetime<![CDATA[]]><![CDATA[Microsoft Teams Link ]]><![CDATA[Ph.D. Dissertation Defense - Eloy Geenjaar]]>28475Title: A data-driven machine learning framework to uncover whole-brain dynamical systems and multimodal interactions in neuroimaging data
Committee:
Dr. Vince Calhoun, ECE, Chair, Advisor
Dr. Christopher Rozell, ECE
Dr. Michael Borich, Emory
Dr. Hannah Choi, Math
Dr. Sergey Plis, GSU
]]> Daniela Staiculescu117852760352026-07-28 22:00:3517852761012026-07-28 22:01:4100eventNeuroimaging research is starting to move from linear machine learning methods towards neural networks. Since neural networks can fit more flexible functions to data, they are a prime candidate for finding relationships that linear machine learning methods are unable to find. However, the most successful applications of data-driven linear machine learning methods, like independent component analysis, to neuroimaging data used neurophysiological priors about whole-brain spatiotemporal dynamics in order to capture interpretable features that provided new insights about the brain. I propose a framework in this dissertation that leverages the flexibility of deep generative models, which are data-driven unsupervised neural networks, and imbues them with more complex neurophysiological priors through inductive biases. The main two axes of neurophysiological priors I explore in this work are the incorporation of dynamics and multimodal information into deep generative models for neuroimaging. My goal is to show that by incorporating neurophysiological priors through inductive biases in deep generative models, the resulting low-dimensional representations are neurophysiologically informed and can be used to make inferences about dynamical or multimodal neuroimaging features. Throughout this work I also focus on the development of interpretability techniques that allow the structure of the low-dimensional representations and the features they capture to be visualized and used to understand what particular spatiotemporal patterns are implicated in cognitive processes or clinical populations. The framework I propose is evaluated on various task-based, resting-state, naturalistic, and clinical fMRI datasets in order to show the validity of both the neurophysiological priors and the framework as a whole. In many cases, I show that the framework is able to uncover either dynamically-informed or multimodally-informed brain patterns that are related to either behavioral variables and/or schizophrenia. Hence, I show that the proposed framework can be used to study cognitive processes, and has the potential to support biomarker development for a wider range of neuropsychiatric disorders.]]>2026-08-06T12:00:00-04:002026-08-06T14:00:00-04:002026-08-06T14:00:00-04:002026-08-06 16:00:002026-08-06 18:00:002026-08-06 18:00:002026-08-06T12:00:00-04:002026-08-06T14:00:00-04:00America/New_YorkAmerica/New_Yorkdatetime2026-08-06 12:00:002026-08-06 02:00:00America/New_YorkAmerica/New_Yorkdatetime<![CDATA[]]><![CDATA[Zoom Link ]]>