University of Cambridge > Talks.cam > ML@CL Seminar Series > Generalized Energy Based Models

Generalized Energy Based Models

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

If you have a question about this talk, please contact Francisco Vargas.

Paper:

Talk is based on this paper.

Abstract:

We introduce the Generalized Energy Based Model (GEBM) for generative modelling. These models combine two trained components: a base distribution (generally an implicit model), which can learn the support of data with low intrinsic dimension in a high dimensional space; and an energy function, to refine the probability mass on the learned support. Both the energy function and base jointly constitute the final model, unlike GANs, which retain only the base distribution (the “generator”). GEB Ms are trained by alternating between learning the energy and the base. We show that both training stages are well-defined: the energy is learned by maximising a generalized likelihood, and the resulting energy-based loss provides informative gradients for learning the base. Samples from the posterior on the latent space of the trained model can be obtained via MCMC , thus finding regions in this space that produce better quality samples. Empirically, the GEBM samples on image-generation tasks are of much better quality than those from the learned generator alone, indicating that all else being equal, the GEBM will outperform a GAN of the same complexity. When using normalizing flows as base measures, GEB Ms succeed on density modelling tasks, returning comparable performance to direct maximum likelihood of the same networks.

About the Speaker:

TBC

Website: https://michaelarbel.github.io/

Part of ML@CL Seminar Series focusing on early career researchers in topics relevant to machine learning and statistics.

This talk is part of the ML@CL Seminar Series series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

© 2006-2021 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity