Tractable loss function and color image generation of multinary restricted Boltzmann machine

TL;DR

This paper introduces differentiable loss functions for multinary RBMs, enhancing their learnability and enabling the generation of colored face images, thus improving their practicality as generative models.

Contribution

It derives new differentiable loss functions for multinary RBMs and demonstrates their effectiveness in generating colored face images.

Findings

Successfully derived differentiable loss functions for multinary RBMs

Demonstrated improved learnability of RBMs with the new loss functions

Generated colored face images using the proposed methods

Abstract

The restricted Boltzmann machine (RBM) is a representative generative model based on the concept of statistical mechanics. In spite of the strong merit of interpretability, unavailability of backpropagation makes it less competitive than other generative models. Here we derive differentiable loss functions for both binary and multinary RBMs. Then we demonstrate their learnability and performance by generating colored face images.