Glancing Transformer for Non-Autoregressive Neural Machine Translation

TL;DR

This paper introduces the Glancing Transformer (GLAT), a non-autoregressive translation model that achieves high-quality results with significantly increased speed by learning word interdependencies for single-pass decoding.

Contribution

The paper proposes GLAT, a novel non-autoregressive model that outperforms previous methods and approaches Transformer quality with only one decoding pass.

Findings

GLAT achieves 8-15x speedup over autoregressive models.

GLAT outperforms previous single-pass NAT methods.

GLAT's translation quality is nearly comparable to Transformer.

Abstract

Recent work on non-autoregressive neural machine translation (NAT) aims at improving the efficiency by parallel decoding without sacrificing the quality. However, existing NAT methods are either inferior to Transformer or require multiple decoding passes, leading to reduced speedup. We propose the Glancing Language Model (GLM), a method to learn word interdependency for single-pass parallel generation models. With GLM, we develop Glancing Transformer (GLAT) for machine translation. With only single-pass parallel decoding, GLAT is able to generate high-quality translation with 8-15 times speedup. Experiments on multiple WMT language directions show that GLAT outperforms all previous single pass non-autoregressive methods, and is nearly comparable to Transformer, reducing the gap to 0.25-0.9 BLEU points.