Past Fortnightly
Open Datasets and Deep Learning Interatomic Potentials for Catalyst Discovery with Zack Ulissi
Abstract:
Machine learning accelerated catalyst discovery efforts have seen much progress in the last few years. Datasets of computational calculations have improved, models to connect surface structure with electronic structure or adsorption energies have gotten more sophisticated, and active learning exploration strategies are becoming routine in discovery efforts.
Large-scale data and machine learning modeling efforts like the Open Catalyst Project have elevated computational catalysis to a first-class problem in the broader machine learning community and led to rapid improvements in accuracy with new state-of-the-art AI/ML models appearing every 3-4 months since 2020. The resulting models are now accurate enough to assist with many day-to-day catalyst simulation efforts (AdsorbML), and to share the capabilities with the community we have released an online demo and representative case studies.
Finally, we’ll discuss current and future efforts to build similar approaches for other climate-related and materials challenges like direct air capture via the OpenDAC collaboration.
Newsletter Subscription
Subscribe to our newsletter and stay updated with the latest in material informatics.
Guest bio
Zack Ulissi
Research Scientist, Meta Fundamental AI Research, Adjunct Professor of Chemical Engineering, CMU
Zack Ulissi is a research scientist on the FAIR Chemistry team at Meta’s Fundamental AI Research lab working on AI/ML for chemistry and climate and on professional leave from Carnegie Mellon University. Prior to Meta, he joined the Department of Chemical Engineering at CMU in 2017 as an Assistant and then Associate Professor. He completed his undergraduate work at the University of Delaware, M.A.St. at Cambridge, PhD in carbon nanotube device modeling at MIT with Michael Strano and Richard Braatz, and post-doc at Stanford in computational catalysis with Jens Nørskov.
