Learning Representations by Maximizing Mutual Information in Variational Autoencoders
This work addresses the limitation of VAEs for representation learning, which is important for researchers in unsupervised learning and generative models, though it appears incremental as it builds on existing VAE methods.
The paper tackles the problem of variational autoencoders (VAEs) not being useful for representation learning by proposing InfoMax-VAE, which combines VAEs with mutual information maximization to enhance amortized inference and improve learned representations, showing it outperforms contemporary approaches like Info-VAE and β-VAE in experiments.
Variational autoencoders (VAEs) have ushered in a new era of unsupervised learning methods for complex distributions. Although these techniques are elegant in their approach, they are typically not useful for representation learning. In this work, we propose a simple yet powerful class of VAEs that simultaneously result in meaningful learned representations. Our solution is to combine traditional VAEs with mutual information maximization, with the goal to enhance amortized inference in VAEs using Information Theoretic techniques. We call this approach InfoMax-VAE, and such an approach can significantly boost the quality of learned high-level representations. We realize this through the explicit maximization of information measures associated with the representation. Using extensive experiments on varied datasets and setups, we show that InfoMax-VAE outperforms contemporary popular approaches, including Info-VAE and $β$-VAE.