Tractable loss function and color image generation of multinary restricted Boltzmann machine
This work provides a method to train RBMs using backpropagation, potentially making them more competitive with other generative models for researchers and practitioners interested in interpretable generative models.
This paper addresses the challenge of training Restricted Boltzmann Machines (RBMs) by deriving differentiable loss functions for both binary and multinary RBMs. The authors demonstrate the learnability and performance of these new loss functions by generating colored face images.
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.