GRADUATE STUDENT SEMINAR, Pinchen Xie

Ab initio generalized Langevin equation: from quantum mechanics to mesoscopic physics

We propose an approach for learning accurately the dynamics of slow, mesoscopic variables from atomistic data obtained from ab-initio quantum mechanical theory, using generalized Langevin equations. The approach is hence called ab initio generalized Langevin equation (AIGLE) [1]. Combined with Deep Potential Molecular Dynamics (developed by PACM alumni Linfeng and Jiequn) and electronic density functional theory, AIGLE opens the door to carrying out first-principles multi-scale modeling for a variety of systems. 

In this talk, I will demonstrate the capability of AIGLE with a simple case study of the mesoscopic domain wall dynamics in a realistic solid state material, based on a microscopic model previously trained on quantum mechanical data in [2]. I will hide the physical details in a black box. So pre-knowledge about quantum mechanics or material science is helpful but not necessary. 

[1]: Xie, P., Car, R. and E, W., 2022. Ab Initio Generalized Langevin Equations. arXiv preprint arXiv:2211.06558.

[2]: Xie, P., Chen, Y., E, W. and Car, R., 2022. Ab initio multi-scale modeling of ferroelectrics: The case of PbTiO3. arXiv preprint arXiv:2205.11839.