Exponential Tilting of Generative Models: Improving Sample Quality by Training and Sampling from Latent Energy

By Zhisheng Xiao, Qing Yan, Yali Amit. 2020

8 citations recorded by the source corpus.

In this paper, we present a general method that can improve the sample quality of pre-trained likelihood based generative models. Our method constructs an energy function on the latent variable space that yields an energy function on samples produced by the pre-trained generative model. The energy based model is efficiently trained by maximizing the data likelihood, and after training, new samples in the latent space are generated from the energy based model and passed through the generator to producing samples in observation space. We show that using our proposed method, we can greatly improve the sample quality of popular likelihood based generative models, such as normalizing flows and VAEs, with very little computational overhead.

Subfields: Computer Graphics and Computer-Aided Design, Computer Vision and Pattern Recognition, Generative Models, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Graphics and Computer-Aided Design

In Computer Vision and Pattern Recognition

In Generative Models

In Signal Processing

Citations in the corpus

VAEBM: A Symbiosis between Variational Autoencoders and Energy-based Models (2020, methodology): “Xiao et al. (2020); Nijkamp et al. (2020) use normalizing flows that transport complex data to latent variables to facilitate MCMC sampling (Hoffman et al., 2019), and thus, their methods can be viewed as EBMs with flow component.”