IBM-Illinois Discovery Accelerator Institute

 Abstract Building new molecular matter is the foundational challenge in chemistry and the confluence of two tools, nonequilibrium control and machine learning, is providing new opportunities to expand our capabilities for both design and control. In this talk, I will describe two efforts that sit at the interface between chemistry and machine learning to shed light on how we can leverage chemical insight to inform large-scale models for molecular design. First, I will discuss adapting "foundation models" with experimental data to inform the design of both proteins and small molecules. In the second part, I will describe how external, nonequilibrium controls can push biomaterial design beyond its static limits. 

Biography Grant Rotskoff is an Assistant Professor of Chemistry at Stanford. His research group focuses on developing robust machine learning approaches that both improve with increasing scale but also respect the physical and chemical constraints that dictate utility. His work applies techniques at the interface of nonequilibrium statistical mechanics, biophysical simulation, and applied mathematics to solve challenging molecular design problems that incorporate multiple length and time scales. Prior to his current position, Grant was a James S. McDonnell Fellow working on machine learning theory at the Courant Institute of Mathematical Sciences at New York University. He completed his Ph.D. at the University of California, Berkeley in the Biophysics graduate group supported by an NSF Graduate Research Fellowship. Grant received an S.B. in Mathematics from the University of Chicago. He is a recipient of the Department of Energy Early Career Award, the National Science Foundation CAREER award, and the Google Research Scholar Award.