Risi Kondor has received the 2016 Defense Advanced Research Projects Agency (DARPA) Young Faculty Award for his proposal, Multiresolution Machine Learning for Molecular Modeling.
Kondor is one of the pioneers of a new approach to molecular dynamics that involves using machine learning to combine classical and quantum mechanical simulations methods, and has been a core participant at a semester long program on this subject at the Institute for Pure and Applied Mathematics at UCLA.
“The classical and quantum mechanical approaches to simulating atomic systems, such as solids, proteins, etc. are really complimentary. While the classical simulations are fast, they completely neglect quantum effects. Quantum simulations are much more accurate, but are far too costly to realistically simulate any system composed of more than a handful of atoms. The idea in the new field of ML/MD is to use Machine Learning to learn the forces that atoms exert on each other from a relatively small number of quantum simulations, and then use these learned force fields in a fast classical simulation. The hope is that this will bring unprecedented accuracy to molecular dynamics and allow us to study phenomena which were previously beyond our reach.” - Assistant Professor Risi Kondor
The goal of the Young Faculty Award is to support high-risk, ground breaking research in a small number of focus areas. Kondor’s award totals $500,000 - to be spent over a two year period.