Abstract
I will start by describing an experiment we did at STOC, which is a top theory conference, where we used advanced Gemini-based AI models to provide pre-submission feedback to authors:
https://research.google/blog/gemini-provides-automated-feedback-for-theoretical-computer-scientists-at-stoc-2026/
The tools we developed for that experiment serve as strong mathematical verifiers and were subsequently used at ICML and NeurIPS. I will next present a collection of case studies demonstrating how researchers have successfully collaborated with variants of such models to solve research open problems, refute conjectures, and generate new proofs across diverse areas in theoretical computer science, as well as other areas such as economics, optimization, and physics. This is based on a corresponding paper (https://arxiv.org/abs/2602.03837) with around 20 testimonials of open problems we made progress on.

Biography:
David Woodruff is a professor at Carnegie Mellon University in the Computer Science Department and a researcher at Google Research. Before that he was a research scientist at IBM Almaden for ten years. He received his PhD from MIT in 2007. His research interests include machine learning, numerical linear algebra, optimization and randomized algorithms. He is the recipient of the 2020 Simons Investigator Award, the 2014 Presburger Award, Best Paper Awards at STOC 2013, PODS 2010, PODS, 2020, PODS 2026, and a STOC 2023 Test of Time Award. At IBM he was a member of the Academy of Technology and a Master Inventor.