Fast Structured Decoding for Sequence Models

TL;DR

This paper introduces a structured inference module using an efficient CRF approximation for non-autoregressive sequence models, significantly improving translation accuracy with minimal latency increase.

Contribution

It proposes a novel structured decoding method incorporating a dynamic transition CRF to enhance non-autoregressive models' consistency and performance.

Findings

Achieves higher BLEU scores than previous non-autoregressive models.

Increases inference speed by only 8-14ms.

Outperforms prior non-autoregressive baselines on multiple datasets.

Abstract

Autoregressive sequence models achieve state-of-the-art performance in domains like machine translation. However, due to the autoregressive factorization nature, these models suffer from heavy latency during inference. Recently, non-autoregressive sequence models were proposed to reduce the inference time. However, these models assume that the decoding process of each token is conditionally independent of others. Such a generation process sometimes makes the output sentence inconsistent, and thus the learned non-autoregressive models could only achieve inferior accuracy compared to their autoregressive counterparts. To improve then decoding consistency and reduce the inference cost at the same time, we propose to incorporate a structured inference module into the non-autoregressive models. Specifically, we design an efficient approximation for Conditional Random Fields (CRF) for non-autoregressive sequence models, and further propose a dynamic transition technique to model positional contexts in the CRF. Experiments in machine translation show that while increasing little latency (8~14ms), our model could achieve significantly better translation performance than previous non-autoregressive models on different translation datasets. In particular, for the WMT14 En-De dataset, our model obtains a BLEU score of 26.80, which largely outperforms the previous non-autoregressive baselines and is only 0.61 lower in BLEU than purely autoregressive models.