Practical Lossless Compression with Latent Variables using Bits Back Coding

TL;DR

This paper introduces BB-ANS, a practical scheme for lossless data compression using latent variable models, demonstrating superior performance on MNIST with potential for further improvements.

Contribution

The paper presents BB-ANS, a near-optimal lossless compression method leveraging latent variable models, and demonstrates its effectiveness with a variational auto-encoder on MNIST.

Findings

Achieved better compression rates than standard methods on MNIST.

The scheme is highly parallelizable, enabling efficient implementation.

Open source implementation is provided for reproducibility.

Abstract

Deep latent variable models have seen recent success in many data domains. Lossless compression is an application of these models which, despite having the potential to be highly useful, has yet to be implemented in a practical manner. We present `Bits Back with ANS' (BB-ANS), a scheme to perform lossless compression with latent variable models at a near optimal rate. We demonstrate this scheme by using it to compress the MNIST dataset with a variational auto-encoder model (VAE), achieving compression rates superior to standard methods with only a simple VAE. Given that the scheme is highly amenable to parallelization, we conclude that with a sufficiently high quality generative model this scheme could be used to achieve substantial improvements in compression rate with acceptable running time. We make our implementation available open source at https://github.com/bits-back/bits-back .