University of Cambridge > Talks.cam > Machine Learning Reading Group @ CUED > Boltzmann Generators and Stochastic Normalizing Flows

Boltzmann Generators and Stochastic Normalizing Flows

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Robert Pinsler.

In statistical physics, systems are modeled by a probability distribution, i.e. the Boltzmann distribution. Being able to draw samples from it is of utmost importance for many applications such as analyzing physical processes as well as drug discovery. Current state of the art sampling procedures such as molecular dynamics simulation and Markov Chain Monte Carlo methods can only generate sequences of highly correlated samples, leaving large parts of the potential landscape unexplored. In this talk, we will present a paper introducing Boltzmann generators which is a method to approximate Boltzmann distributions with a normalizing flow in order to draw independent samples from it. Furthermore, we will discuss a follow up work about a novel flow based model, called Stochastic Normalizing Flows, which is supposedly, among other merits, even better suited to draw independent samples from Boltzmann distributions.

This talk is part of the Machine Learning Reading Group @ CUED series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

© 2006-2020 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity