Oct. 10, 2024 - Simon Batzner: Materials Discovery Using Simulations and Deep Learning

Despite the recent advances in physical simulations and machine learning, the exploration of novel inorganic crystals remains constrained by the expensive trial-and-error approaches. Recent developments in deep learning have shown that models can attain emergent predictive capabilities with increasing data and computation, in fields such as language, vision, and biology. In this talk, we will present our recent results on how graph networks trained at scale can reach unprecedented levels of generalization, improving the efficiency of materials discovery by an order of magnitude. We will show how the scale and diversity unlock surprising modeling capabilities for downstream applications, including predictions of crystalline stability, ionic conduction, and structural properties of amorphous materials.

 

Bio: Simon Batzner is a Research Scientist at Google DeepMind. His research interests lie in large-scale deep learning and molecular and materials discovery. Prior to joining DeepMind, he worked with Boris Kozinksy on building deep learning systems for materials simulations, in particular on learning symmetry-preserving representations of geometric structures. His Ph.D. research was a finalist  for the 2023 ACM Gordon Bell Prize, also known as the "Nobel Prize of Supercomputing". Simon holds a Ph.D. in Applied Mathematics from Harvard University and a Master's from MIT.