Self-Adapting Noise-Contrastive Estimation for Energy-Based Models

TL;DR

This paper introduces a self-adapting noise-contrastive estimation method for energy-based models that eliminates the need for an auxiliary noise model by using static EBM instances during training, improving synthesis quality.

Contribution

It proposes a novel self-adapting NCE algorithm that uses static EBM instances as noise, simplifying training and enhancing performance without extra memory overhead.

Findings

Shorter noise update intervals lead to higher synthesis quality.

The method generalizes maximum likelihood learning for EBMs.

Experimental results validate the effectiveness of the approach.

Abstract

Training energy-based models (EBMs) with noise-contrastive estimation (NCE) is theoretically feasible but practically challenging. Effective learning requires the noise distribution to be approximately similar to the target distribution, especially in high-dimensional domains. Previous works have explored modelling the noise distribution as a separate generative model, and then concurrently training this noise model with the EBM. While this method allows for more effective noise-contrastive estimation, it comes at the cost of extra memory and training complexity. Instead, this thesis proposes a self-adapting NCE algorithm which uses static instances of the EBM along its training trajectory as the noise distribution. During training, these static instances progressively converge to the target distribution, thereby circumventing the need to simultaneously train an auxiliary noise model. Moreover, we express this self-adapting NCE algorithm in the framework of Bregman divergences and show that it is a generalization of maximum likelihood learning for EBMs. The performance of our algorithm is evaluated across a range of noise update intervals, and experimental results show that shorter update intervals are conducive to higher synthesis quality.