Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore advanced generative modeling techniques for materials discovery in this conference talk from IPAM's seminar series on electrochemistry modeling. Learn about the Open Materials Generation (OMatG) framework, which uses stochastic interpolants for the generative design and discovery of inorganic crystalline materials. Discover how computational approaches can effectively learn the manifold of stable crystal structures within infinite design spaces while relying on robust benchmarks and minimal, information-rich datasets for meaningful evaluation. Examine the framework's performance in two critical tasks: crystal structure prediction (CSP) for given compositions and de novo generation (DNG) for discovering entirely novel, stable, and unique structures. Understand how revised and extended common metrics and datasets demonstrate OMatG's state-of-the-art performance, surpassing purely flow- and diffusion-based implementations. Gain insights into the importance of flexible deep learning frameworks and sensible benchmarks for accelerating progress in materials science and enabling technological advancements through new material discovery.
