Neural Joint Source-Channel Coding

TL;DR

This paper introduces a neural joint source-channel coding method using a discrete variational autoencoder that learns to compress and error-correct data simultaneously, outperforming traditional schemes and enabling fast decoding.

Contribution

It proposes a novel neural joint source-channel coding framework with a discrete VAE that improves compression, error correction, and decoding speed over classical methods.

Findings

Codes are competitive with separation schemes.

Learned representations are robust for downstream tasks.

Neural decoder is nearly ten times faster than traditional methods.

Abstract

For reliable transmission across a noisy communication channel, classical results from information theory show that it is asymptotically optimal to separate out the source and channel coding processes. However, this decomposition can fall short in the finite bit-length regime, as it requires non-trivial tuning of hand-crafted codes and assumes infinite computational power for decoding. In this work, we propose to jointly learn the encoding and decoding processes using a new discrete variational autoencoder model. By adding noise into the latent codes to simulate the channel during training, we learn to both compress and error-correct given a fixed bit-length and computational budget. We obtain codes that are not only competitive against several separation schemes, but also learn useful robust representations of the data for downstream tasks such as classification. Finally, inference amortization yields an extremely fast neural decoder, almost an order of magnitude faster compared to standard decoding methods based on iterative belief propagation.