Learning the Irreducible Representations of Commutative Lie Groups

TL;DR

This paper introduces a probabilistic model for compact commutative Lie groups that learns invariant and disentangled data representations, enabling effective classification through fully tractable Bayesian inference.

Contribution

It presents a novel Bayesian conjugacy relation for compact commutative Lie groups, facilitating tractable inference and improved invariant representation learning.

Findings

Learned invariant representations are highly effective for classification.

Model successfully captures symmetries like rotation and cyclic translation.

Provides a new approach to disentangling data based on physical principles.

Abstract

We present a new probabilistic model of compact commutative Lie groups that produces invariant-equivariant and disentangled representations of data. To define the notion of disentangling, we borrow a fundamental principle from physics that is used to derive the elementary particles of a system from its symmetries. Our model employs a newfound Bayesian conjugacy relation that enables fully tractable probabilistic inference over compact commutative Lie groups -- a class that includes the groups that describe the rotation and cyclic translation of images. We train the model on pairs of transformed image patches, and show that the learned invariant representation is highly effective for classification.