HSIC-InfoGAN: Learning Unsupervised Disentangled Representations by Maximising Approximated Mutual Information

TL;DR

This paper introduces HSIC-InfoGAN, a method that uses the Hilbert-Schmidt Independence Criterion to learn disentangled representations without auxiliary networks, potentially benefiting medical image analysis.

Contribution

It proposes replacing the mutual information maximization in InfoGAN with HSIC, simplifying the model and reducing training complexity.

Findings

HSIC-InfoGAN achieves comparable disentanglement to traditional InfoGAN.

The method eliminates the need for an auxiliary network.

Potential applications in medical imaging are discussed.

Abstract

Learning disentangled representations requires either supervision or the introduction of specific model designs and learning constraints as biases. InfoGAN is a popular disentanglement framework that learns unsupervised disentangled representations by maximising the mutual information between latent representations and their corresponding generated images. Maximisation of mutual information is achieved by introducing an auxiliary network and training with a latent regression loss. In this short exploratory paper, we study the use of the Hilbert-Schmidt Independence Criterion (HSIC) to approximate mutual information between latent representation and image, termed HSIC-InfoGAN. Directly optimising the HSIC loss avoids the need for an additional auxiliary network. We qualitatively compare the level of disentanglement in each model, suggest a strategy to tune the hyperparameters of HSIC-InfoGAN, and discuss the potential of HSIC-InfoGAN for medical applications.