Context-Aware Monolingual Repair for Neural Machine Translation

TL;DR

This paper introduces a monolingual post-editing model called DocRepair that improves the consistency of sentence translations in context, using only target language data, and demonstrates significant improvements in translation quality and coherence.

Contribution

The paper presents a novel monolingual sequence-to-sequence model for post-editing translations to ensure contextual consistency, trained solely on target language data.

Findings

Large improvements in contextual translation phenomena

Enhanced BLEU scores for English-Russian translation

Human evaluators prefer corrected translations

Abstract

Modern sentence-level NMT systems often produce plausible translations of isolated sentences. However, when put in context, these translations may end up being inconsistent with each other. We propose a monolingual DocRepair model to correct inconsistencies between sentence-level translations. DocRepair performs automatic post-editing on a sequence of sentence-level translations, refining translations of sentences in context of each other. For training, the DocRepair model requires only monolingual document-level data in the target language. It is trained as a monolingual sequence-to-sequence model that maps inconsistent groups of sentences into consistent ones. The consistent groups come from the original training data; the inconsistent groups are obtained by sampling round-trip translations for each isolated sentence. We show that this approach successfully imitates inconsistencies we aim to fix: using contrastive evaluation, we show large improvements in the translation of several contextual phenomena in an English-Russian translation task, as well as improvements in the BLEU score. We also conduct a human evaluation and show a strong preference of the annotators to corrected translations over the baseline ones. Moreover, we analyze which discourse phenomena are hard to capture using monolingual data only.