Disentangling representations in Restricted Boltzmann Machines without adversaries
This work addresses the difficulty of adversarial training in disentangling representations, offering a simpler alternative for researchers in unsupervised learning, though it appears incremental as it builds on existing RBM frameworks.
The authors tackled the problem of disentangling representations in unsupervised machine learning by proposing a method that avoids adversarial training, applying it to Restricted Boltzmann Machines and demonstrating effectiveness on datasets like CelebA, MNIST, and protein families, with analytical computation of the log-likelihood cost.
A goal of unsupervised machine learning is to build representations of complex high-dimensional data, with simple relations to their properties. Such disentangled representations make easier to interpret the significant latent factors of variation in the data, as well as to generate new data with desirable features. Methods for disentangling representations often rely on an adversarial scheme, in which representations are tuned to avoid discriminators from being able to reconstruct information about the data properties (labels). Unfortunately adversarial training is generally difficult to implement in practice. Here we propose a simple, effective way of disentangling representations without any need to train adversarial discriminators, and apply our approach to Restricted Boltzmann Machines (RBM), one of the simplest representation-based generative models. Our approach relies on the introduction of adequate constraints on the weights during training, which allows us to concentrate information about labels on a small subset of latent variables. The effectiveness of the approach is illustrated with four examples: the CelebA dataset of facial images, the two-dimensional Ising model, the MNIST dataset of handwritten digits, and the taxonomy of protein families. In addition, we show how our framework allows for analytically computing the cost, in terms of log-likelihood of the data, associated to the disentanglement of their representations.