|
Katherine Wu I'm a PhD student in Computer Science at Cornell University, where I am advised by Alexandra Silva. My research lies broadly at the intersection of probabilistic reasoning in AI and formal methods. I am particularly interested in probabilistic programming, a programming paradigm centered on designing languages and techniques that let users express uncertain models as programs and automatically reason about their behavior, and in making probabilistic inference more scalable, expressive, and reliable. I am grateful to be supported by an NSF Graduate Research Fellowship. This summer 2026, I have also been exploring differentiable programming as a way to bring gradient-based optimization to probabilistic programs, as an intern at Argonne with Jan Hückelheim. Previously, I graduated from Stony Brook University with a BS in computer science (with honors) and a BS in mathematics. I am fortunate to have worked on several research projects while there: with Jeff Heinz on grammar induction, with Erez Zadok on formal semantics for serverless computing, and with Annie Liu on question answering with logic programming. |
|
Publications |
|
Type-Directed Discretization of Probabilistic Programs
Katherine Wu, Jules Jacobs, Kevin Batz, Alexandra Silva OOPSLA 2026, to appear. We introduce Slice, a type-directed transformation that exactly discretizes recursive, higher-order probabilistic programs with continuous distributions while preserving their semantics, enabling exact inference with discrete backends. |
|
LP-LM: No Hallucinations in Question Answering with Logic Programming
Katherine Wu, Yanhong A. Liu International Conference on Logic Programming (ICLP 2024) We propose a logic system called LP-LM that retrieves verifiable answers from a given knowledge base and evaluate it on several question-answering tasks. |
|
Balancing Costs and Durability for Serverless Data
Alex Merenstein, Xinran Wang, Vasily Tarasov, Prajjawal Agarwal, Scott Guthridge, Kapil Thakkar, Katherine Wu, Ali Anwar, Erez Zadok IEEE Symposium on Massive Storage Systems and Technologies (MSST 2024) We develop a mathematical model and execution framework that balance the cost of storing serverless data durably against the cost of recreating it, reducing storage costs by up to 3× without exceeding baseline costs. |
|
String Extension Learning Despite Noisy Intrusions
Katherine Wu, Jeffrey Heinz International Conference on Grammatical Inference (ICGI 2023) We examine the conditions in which string extension learning algorithms are able to identify classes of formal languages in the limit from noisy data presentations in polynomial time. |
Teaching |
|
CS 4850: Probability, Vectors, and Matrices in Computing
|
|
CS 4820: Introduction to Analysis of Algorithms
|
|
CS 150: Foundations of Computer Science - Honors
|