Boltzmann generators (BGs) [1] are neural network-based importance samplers that can generate configurations of molecular systems. In contrast to classical molecular sampling techniques, BGs generate independent samples from the Boltzmann distribution in one shot…
I will give an overview of Markov State Modeling, deep learning of Markov models via variational approaches, and more recent approaches to efficiently emulate protein thermodynamics and kinetics with generative deep learning methods…