Bounded Information Rate Variational Autoencoders
arXiv:1807.07306
Abstract
This paper introduces a new member of the family of Variational Autoencoders (VAE) that constrains the rate of information transferred by the latent layer. The latent layer is interpreted as a communication channel, the information rate of which is bound by imposing a pre-set signal-to-noise ratio. The new constraint subsumes the mutual information between the input and latent variables, combining naturally with the likelihood objective of the observed data as used in a conventional VAE. The resulting Bounded-Information-Rate Variational Autoencoder (BIR-VAE) provides a meaningful latent representation with an information resolution that can be specified directly in bits by the system designer. The rate constraint can be used to prevent overtraining, and the method naturally facilitates quantisation of the latent variables at the set rate. Our experiments confirm that the BIR-VAE has a meaningful latent representation and that its performance is at least as good as state-of-the-art competing algorithms, but with lower computational complexity.
arXiv commentPresented at KDD 2018 Deep Learning Day. Minor changes and correction of rate calculations, overall results remain the same
References in corpus (10)
- Stochastic Backpropagation and Approximate Inference in Deep Generative Models
- Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks
- Improved Training of Wasserstein GANs
- InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets
- DRAW: A Recurrent Neural Network For Image Generation
- Opening the Black Box of Deep Neural Networks via Information
- Deep Convolutional Inverse Graphics Network
- InfoVAE: Information Maximizing Variational Autoencoders
- Adversarial Autoencoders
- Variational Lossy Autoencoder