题 目: Learning Assisted Molecular Modeling Beyond Deep Potential
报告人: Linfeng Zhang, Beijing Institute of Big Data Research
时 间: 3月15日（周一）13:00-14:00
地 点: 北京大学老化学楼东配楼101报告厅
主持人: 裴剑锋 研究员
Deep Potential (DP) has proven to be an effective way to fit the potential energy surface (PES) data calculated by the density functional theory (DFT), thereby accelerating molecular dynamics with DFT accuracy by several orders of magnitude. Here we consider two issues beyond DP: 1) pushing the limit of accuracy: how to improve the accuracy of DFT itself? 2) pushing the limit of efficiency: given an accurate by rugged PES, how to efficiently sample the configuration space? For the first issue, we introduce Deep Kohn-Sham (DeePKS), our prescription for developing data-driven DFT models with chemical accuracy; for the second issue, we introduce Reinforced Dynamic (RiD), which helps to explore the free energy landscape with more than 100 collective variables. The algorithmic aspects of these schemes will be strengthened with examples.
Linfeng Zhang is temporarily working as a research scientist at the Beijing Institute of Big Data Research. In the May of 2020, he graduated from the Program in Applied and Computational Mathematics (PACM), Princeton University, working with Profs. Roberto Car and Weinan E. Linfeng has been focusing on developing machine learning based physical models for electronic structures, molecular dynamics, as well as enhanced sampling. He is one of the main developers of DeePMD-kit, a very popular deep learning based open-source software for molecular simulation in physics, chemistry, and materials science. He is a recipient of the 2020 ACM Gordon Bell Prize for their project “Pushing the limit of molecular dynamics with ab initio accuracy to 100 million atoms with machine learning”.