PhD Thesis Defense

This thesis develops and analyzes algorithms for several key problems related to eigenvalue and eigenvector estimation.

This dissertation argues for a different approach to sorting and filtering social media feeds.

This research advances methodologies for analyzing content from online datasets, interpreting intent, retrieving knowledge, and leveraging these insights.

In this defense, I outline a program to move from post-hoc explanations toward a principled science of reverse-engineering LLM internals.

This research investigates the design and implementation of ultrasonic sensing systems that allow everyday devices to perceive and interpret signals beyond the

This thesis studies questions from three complementary perspectives for real-world scenarios.

This dissertation argues that how people perceive and interpret bias in AI outputs is shaped by how the outputs are presented and who is interpreting them.

This dissertation investigates how NLP methods can enable scalable educational assessment and AI systems across key components of the educational pipeline.

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.