The Neural Noisy Channel

TL;DR

This paper introduces a neural noisy channel model for sequence transduction tasks, leveraging unpaired data and a latent variable to improve performance over direct models across various NLP tasks.

Contribution

It presents a neural noisy channel decoding framework that effectively incorporates unpaired data and a latent variable for improved sequence transduction.

Findings

Outperforms direct models on summarization, inflection, and translation.

Benefits significantly from unpaired output data.

Uses a tractable beam search decoder with a latent variable.

Abstract

We formulate sequence to sequence transduction as a noisy channel decoding problem and use recurrent neural networks to parameterise the source and channel models. Unlike direct models which can suffer from explaining-away effects during training, noisy channel models must produce outputs that explain their inputs, and their component models can be trained with not only paired training samples but also unpaired samples from the marginal output distribution. Using a latent variable to control how much of the conditioning sequence the channel model needs to read in order to generate a subsequent symbol, we obtain a tractable and effective beam search decoder. Experimental results on abstractive sentence summarisation, morphological inflection, and machine translation show that noisy channel models outperform direct models, and that they significantly benefit from increased amounts of unpaired output data that direct models cannot easily use.