Logical Reasoning with a Touch of AI: Andrew Wu (Amherst College)

The Amherst College Data Science Initiative presents:

Andrew Wu (Amherst College)
Tuesday, 9/23 4pm, Science Center A131 with refreshments

https://www.amherst.edu/about/science_at_amherst/data-science-initiative

Proving the correctness of computer programs is a long-standing challenge in Computer Science, as it requires not only rigorous logical deductions but also high-level understanding of the program structure. Existing automated logical reasoning tools excel at the former but struggle with the latter, leaving many practical problems beyond their reach. In this talk, I overview several research directions, developed over the past two years, that propose methodologies for leveraging the demonstrated code-understanding capability of large language models (LLMs) and combining them with logical reasoning tools for automated program verification. We then take a closer look at one of those methods, Lemur, which demonstrated practical improvements on a suite of synthetic and competition benchmarks.
Bio: Andrew Wu is an Assistant Professor in Computer Science and an affiliated researcher at VMware by Broadcom. He obtained his PhD in Computer Science from Stanford University. Andrew's work aims to bring together automated reasoning and machine learning to create safer and smarter computer systems. He is the lead developer of the Marabou framework, a state-of-the-art neural network verification tool widely used in academia and industry. His work has been published in top venues in formal methods and artificial intelligence.

Light refreshments will be served before the talk.