Physics Colloquium: Trustworthy AI Tools for Science

Physics Colloquium
Trustworthy AI Tools for Science

Dr. Ashley Dale
Eric and Wendy Schmidt AI in Science Postdoctoral Fellow
Dept. of Material Science, University of Toronto

Event Details: As AI/ML pervades the scientific community, physicists need to trust that these models and methods are human aligned while appropriately obeying fundamental laws of nature. My recent work seeks trustworthiness for scientific AI/ML along three dimensions: philosophical, computational, and experimental. Science-oriented trustworthy AI frameworks provide a philosophical approach to human alignment in the laboratory; these frameworks evaluate AI/ML for properties such as interpretability and fairness in ways that are meaningful to scientists. To determine whether an AI/ML model obeys physical laws, computational approaches leverage novel explainable uncertainty quantification (xUQ) methods. Finally, because physical models benefit from physical validation, proposed xUQ methods will be evaluated in a self-driving laboratory for the discovery of novel ferroelectric materials. In this talk, I present progress towards these goals and look ahead to how trustworthy AI/ML can support physicists in their future research.

Bio: Ashley Dale is an Eric and Wendy Schmidt AI in Science Postdoctoral Fellow in the Department of Materials Science and Engineering at the University of Toronto, with affiliations at Natural Resources Canada CanmetMATERIALS and The Vector Institute. She earned her Ph.D. in Electrical and Computer Engineering from Purdue University in Indianapolis in 2024, and her Ph.D. in Physics from Indiana University Indianapolis in 2024. In 2025, she was named a Trailblazer in Engineering fellow by Purdue University. Her research focuses on developing methods to evaluate the trustworthiness of scientific AI/ML, with applications in self-driving laboratories and novel ferroelectric materials for next generation computational devices. Her work sits at the intersection of experimental physics, materials discovery, cyber-physical systems, and trustworthy AI methods.

Intended Audience:
All are Welcome!

To request an interpreter, please visit myaccess.rit.edu