Theoretical Physics for Deep Learning

Organizers:

*Maissam Barkeshli, University of Maryland
Andrey Gromov, Brown University
**Alexander Maloney, McGill University
Dan Roberts, Massachusetts Institute of Technology
Eva Silverstein, Stanford University
**James Sully, Anthropic
**Sho Yaida, Meta AI

The rapid growth in the popularity of deep learning has been fueled by transformational advances in the capabilities of artificial intelligence. There is also considerable interest in applications of machine learning and optimization dynamics to a wide variety of scientific and mathematical problems, including data analysis, novel numerical approaches to partial differential equations, wavefunction approximation in condensed matter and AMO physics, and the discovery and verification of mathematical theorems. We invite applications from scientists interested in research at the intersection of physics and artificial intelligence, and in particular on the construction of physics-inspired models of statistical machine learning, the dynamics of optimization, and the workings of deep neural networks. We welcome applications from across the physics community, including high energy, condensed matter and statistical physicists, as well as computer scientists, machine learning researchers, and mathematicians.

*organizer responsible for participant diversity
**scientific advisor