Modeling Coverage for Neural Machine Translation

TL;DR

This paper introduces a coverage mechanism in neural machine translation to track attention history, reducing translation errors and improving both translation and alignment quality.

Contribution

It proposes a novel coverage-based NMT model that incorporates a coverage vector into the attention mechanism to better handle untranslated words.

Findings

Significant improvement in translation quality.

Enhanced alignment accuracy.

Reduction in over- and under-translation issues.

Abstract

Attention mechanism has enhanced state-of-the-art Neural Machine Translation (NMT) by jointly learning to align and translate. It tends to ignore past alignment information, however, which often leads to over-translation and under-translation. To address this problem, we propose coverage-based NMT in this paper. We maintain a coverage vector to keep track of the attention history. The coverage vector is fed to the attention model to help adjust future attention, which lets NMT system to consider more about untranslated source words. Experiments show that the proposed approach significantly improves both translation quality and alignment quality over standard attention-based NMT.