PhD Thesis Defense

I develop a Gumbel random process-based approach that improves the optimization  landscape, resulting in a more stable and expressive variant of Box Embedding.

This thesis significantly generalizes the scope of metareasoning.

This thesis develops scalable, accurate tools and benchmarks for factuality assessment that set higher standards for AI evaluation.

This thesis develops and analyzes principled, efficient, and scalable algorithms for answering marginal queries.

This dissertation addresses challenges to encrypted databases through the following contributions: obliviousness and query expressiveness.

This thesis advances fine-grained reasoning under limited human supervision through several complementary approaches.

This dissertation examines techniques for learning structure and discusses how the resulting background knowledge is used in decision making under uncertainty.

This dissertation explores how low-overhead dynamic analysis can be used to improve the reliability of Python software.

This dissertation develops a comprehensive framework for high-fidelity motion characterization, addressing challenges of sensitivity, robustness, and privacy.

This dissertation advances the design of practical privacy-preserving machine learning systems.

This thesis develops control frameworks for legged robots that bridge model-based trajectory optimization and data-driven learning

 This thesis examines three complementary roles of motion in computer vision.