Research Interests: I am broadly interested in advancing principled ML methodology. My current focus is on modeling complex dynamical systems (e.g., molecular dynamics, reinforcement learning, etc.), and application of generative models such as inverse design. I am particularly interested in coarse-graining and abstraction related approaches.
- 22/04 New preprint: an end-to-end system for learning to simulate coarse-grained molecular dynamics, that runs several orders of magnitude faster than classical force fields! Check it out at ICLR and look forward to seeing you at MRS Spring Meeting on May 8th! Previously presented at APS March Meeting.
- 22/04 I will join Microsoft Research for summer 2022, working with domain experts on material discovery for carbon capture.
- 22/04 Crystal Diffusion Variational Autoencoder for Periodic Material Generation addressing the generation of periodic material structures using a diffusion model decoder, find us at ICLR 2022!