CSMB Colloquium@T²P
-
CSMB Center for the Science of Materials Berlin
Zum Großen Windkanal 2,
12489 Berlin
Room 3.264 and online
Uncertainty-aware molecular dynamics from Bayesian active learning for phase transformations and thermal transport in SiC
Dr. Yu Xie
Microsoft AI4Science Berlin
Abstract:
Machine learning interatomic force fields are promising for combining high computational efficiency and accuracy in modeling quantum interactions and simulating atomistic dynamics. Active learning methods have been developed to train force fields efficiently and automatically. Among them, Bayesian active learning utilizes principled uncertainty quantification to make data acquisition decisions. In this talk, we present our development of a general Bayesian active learning workflow, where the force field is constructed from a sparse Gaussian process regression model based on atomic cluster expansion descriptors. Orders of magnitude of speedup is achieved by the development of an approximate mapping method, and the code implementation of GPU acceleration. We demonstrate the autonomous active learning workflow by training a Bayesian force field model for silicon carbide (SiC) polymorphs in only a few days of computer time and show that pressure-induced phase transformations are accurately captured. The resulting model exhibits close agreement with both ab initio calculations and experimental measurements, and outperforms existing empirical models on vibrational and thermal properties. The active learning workflow readily generalizes to a wide range of material systems and accelerates their computational understanding.
Short Bio:
Dr. Yu Xie is a senior researcher at Microsoft AI for Science in Berlin. She obtained PhD from Harvard University, working with Prof. Boris Kozinsky, with a focus of development of machine learning force field and Bayesian active learning using kernel-based method, and its applications in molecular dynamics simulations of phase transitions of materials.
The talk will also be broadcast via Zoom:
Zoom Link: https://hu-berlin.zoom-x.de/j/63972785525?pwd=oJlOzIouJmLO5x5QLxIsyoIsXC3tbb.1
Meeting ID: 639 7278 5525
Password: 676366