Jay Wadekar (UT Austin): Schmidt AI in Science Speaker Series
Organized by the University of Chicago’s Eric and Wendy Schmidt AI in Science Fellowship Program.
Agenda
3:45pm – 4:00pm: Light refreshments available
4:00pm – 4:45pm: Presentation
4:45pm – 5:00pm: Q&A
Meeting Location
Faraco Family Commons, Room 105
Data Science Institute
5460 S University Ave
Title: Agentic AI for Scientific Discovery: From Gravitational-Wave Searches to Analytic Equations
Abstract: I will show how LLM coding agents can turn quantitative feedback from simulations and independent verification into scientific discoveries. I will present examples spanning gravitational-wave astronomy, cosmology, and nonlinear water waves. In gravitational-wave searches, agents optimized a detection pipeline using feedback from simulated signal injections, increasing its sensitive volume-time for high-mass binary black hole mergers by about 50% and enabling the detection of new transients in public data. Guided by physical ansätze, agents also constructed compact analytic waveform models with an eightfold speedup and accuracy exceeding symbolic regression and machine learning baselines. In cosmology, agent-discovered equations generalized across discrepant simulation suites with only one simulator-dependent coefficient. In water-wave scattering, separating construction from independent verification enabled a multi-agent team to discover a new closed-form six-wave scattering amplitude. I will conclude with limitations of current agents and lessons for building reliable scientific workflows.
Bio: Jay Wadekar is an assistant professor of physics at the University of Texas at Austin, in the Center for Gravitational Physics, and an affiliated faculty member at the Oden Institute. His group builds agentic AI systems that must justify their results against data and established physics. His research spans agentic AI, gravitational-wave astrophysics, cosmology, and interpretable machine learning. He currently co-chairs NASA’s AI/ML Science and Technology Interest Group. Previously, he was a postdoctoral fellow at Johns Hopkins University and a member of the Institute for Advanced Study. He received his PhD from New York University and his B.Tech. from IIT Bombay.
Yuxin Chen (UPenn): Schmidt AI in Science Speaker Series
Armita Nourmohammad (Yale): Schmidt AI in Science Speaker Series
Aly Azeem Khan (UChicago): Schmidt AI in Science Speaker Series