PhD Dissertation Proposal Defense

This dissertation studies communication-efficient multi-device inference for Transformer models under bandwidth-limited settings.

In this thesis, we identify and rectify some remaining limitations of such cryptographic constructions and their proofs of security.

This thesis investigates architectural principles for large-scale quantum networks across three settings.

In the first part of this thesis, we formulate a general-purpose information-theoretic criterion that allows further improving SSL representations.

This thesis investigates generative models for parametric curves and surfaces.

We focus on three fundamental processes: quantum teleportation with noisy memories, error correction in quantum storage, and distributed blind quantum computing

In this thesis, I focus on the computability of allocations which maximize well known fairness objectives

This thesis addresses challenges with robust algorithms tackling different aspects of uncertainty, including RL, resource allocation, explainability and LLMs.

This dissertation shows that two distinct fields - differential privacy and computational ecology - can be addressed through a unified methodological framework.

This dissertation presents a new foundation for flexible and secure replication of blackbox stateful services.

This dissertation addresses how computer science instructors can redesign course structures and pedagogy in response to the widespread availability of LLMs.

In the first part of my thesis, I introduce split parallelism, a novel  abstraction that addresses the limitations of traditional data parallelism on GPUs.