Explore the intersection of artificial intelligence and chemistry in this comprehensive talk by Pratyush Tiwary. Delve into the challenges and opportunities of integrating AI with theoretical and simulation methods in chemistry for new discoveries. Learn about innovative approaches like the Past-future Information Bottleneck, Reweighted Autoencoded Variational Bayes for Enhanced Sampling (RAVE), and Denoising Diffusion Probabilistic Models (DDPM). Discover how these techniques are applied to complex molecular systems, including protein kinases, riboswitches, and crystal polymorph nucleation. Gain insights into the potential of "Artificial Chemical Intelligence" for enabling smart molecular discovery and pushing the boundaries of computational chemistry.
Overview
Syllabus
- Intro
- Overview of Grand Challenges in the Field
- AI + Statistical Mechanics + Molecular Simulations
- Molecular Dynamics - Powerful but Limited
- Past-future Information Bottleneck
- Reweighted Autoencoded Variational Bayes for Enhanced Sampling RAVE and Applications
- Replica Exchange Molecular Dynamics REMD
- Denoising Diffusion Probabilistic Models DDPM
- Replica Exchange at Home
- Conclusion
- Q+A
Taught by
Valence Labs
