Regensburg 2019 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

TUT: Tutorials

TUT 2: Statistical Physics Methods for Data Science in Physics (joint session SOE/TUT/DY)

TUT 2.3: Tutorial

Sunday, March 31, 2019, 17:40–18:30, H3

Multiscale simulations of soft matter augmented by data-driven methods — •Tristan Bereau — Max Planck Institute for Polymer Research

Multiscale simulations, all the way from quantum chemistry to continuum mechanics, probe a variety of length and time scales relevant to soft-matter systems. In this tutorial, I will describe different strategies to help improve physics-based simulations with recently-developed data-driven methods and concepts. Applications discussed will include Bayesian inference for molecular kinetics, machine learning to vastly improve force-field transferability, and high-throughput screening to explore chemical compound space.