Generator Reversal

TL;DR

This paper introduces a novel approach to training generative models by reversing the generator to estimate flexible code distributions, leading to more powerful models and improved latent structure modeling.

Contribution

It proposes a non-parametric method to estimate code distributions by reversing the generator, enhancing model flexibility and control over generalization.

Findings

More powerful generative models achieved

Improved modeling of latent structure

Explicit control of generalization degree

Abstract

We consider the problem of training generative models with deep neural networks as generators, i.e. to map latent codes to data points. Whereas the dominant paradigm combines simple priors over codes with complex deterministic models, we propose instead to use more flexible code distributions. These distributions are estimated non-parametrically by reversing the generator map during training. The benefits include: more powerful generative models, better modeling of latent structure and explicit control of the degree of generalization.