Commutative Lie Group VAE for Disentanglement Learning

Abstract

We view disentanglement learning as discovering an underlying structure that equivariantly reflects the factorized variations shown in data. Traditionally, such a structure is fixed to be a vector space with data variations represented by translations along individual latent dimensions. We argue this simple structure is suboptimal since it requires the model to learn to discard the properties (e.g. different scales of changes, different levels of abstractness) of data variations, which is an extra work than equivariance learning. Instead, we propose to encode the data variations with groups, a structure not only can equivariantly represent variations, but can also be adaptively optimized to preserve the properties of data variations. Considering it is hard to conduct training on group structures, we focus on Lie groups and adopt a parameterization using Lie algebra. Based on the parameterization, some disentanglement learning constraints are naturally derived. A simple model named Commutative Lie Group VAE is introduced to realize the group-based disentanglement learning. Experiments show that our model can effectively learn disentangled representations without supervision, and can achieve state-of-the-art performance without extra constraints.

arXiv commentAccepted in ICML2021

References in corpus (14)

• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets
• Deep Convolutional Inverse Graphics Network
• Disentangling by Factorising
• Deep Learning Face Attributes in the Wild
• Generalizing Convolutional Neural Networks for Equivariance to Lie Groups on Arbitrary Continuous Data
• Learning Hierarchical Features from Generative Models
• Symmetry-Based Disentangled Representation Learning requires Interaction with Environments
• Theory and Evaluation Metrics for Learning Disentangled Representations
• Progressive Learning and Disentanglement of Hierarchical Representations
• Learning Group Structure and Disentangled Representations of Dynamical Environments
• Where and What? Examining Interpretable Disentangled Representations
• Learning Disentangled Representations with Latent Variation Predictability
• Linear Disentangled Representations and Unsupervised Action Estimation