Learning to Disentangle Factors of Variation with Manifold Interaction

Many latent factors of variation interact to generate sensory data; for example, pose, morphology and expression in face images. In this work, we propose to learn manifold coordinates for the relevant factors of variation and to model their joint interaction. Many existing feature learning algorithms focus on a single task and extract features that are sensitive to the task-relevant factors and invariant to all others. However, models that just extract a single set of invariant features do not exploit the relationships among the latent factors. To address this, we propose a higher-order Boltzmann machine that incorporates multiplicative interactions among groups of hidden units that each learn to encode a distinct factor of variation. Furthermore, we propose correspondence-based training strategies that allow effective disentangling. Our model achieves state-of-the-art emotion recognition and face verification performance on the Toronto Face Database. We also demonstrate disentangled features learned on the CMU Multi-PIE dataset.