DEEP: DEnoising Entity Pre-training for Neural Machine Translation

TL;DR

DEEP introduces a denoising pre-training approach leveraging monolingual data and knowledge bases to enhance named entity translation accuracy in neural machine translation models.

Contribution

It proposes a novel denoising entity pre-training method and a multi-task learning strategy to improve named entity translation in NMT.

Findings

Up to 1.3 BLEU score improvement on English-Russian translation.

Up to 9.2 entity accuracy points gained.

Significant improvements over auto-encoding baselines.

Abstract

It has been shown that machine translation models usually generate poor translations for named entities that are infrequent in the training corpus. Earlier named entity translation methods mainly focus on phonetic transliteration, which ignores the sentence context for translation and is limited in domain and language coverage. To address this limitation, we propose DEEP, a DEnoising Entity Pre-training method that leverages large amounts of monolingual data and a knowledge base to improve named entity translation accuracy within sentences. Besides, we investigate a multi-task learning strategy that finetunes a pre-trained neural machine translation model on both entity-augmented monolingual data and parallel data to further improve entity translation. Experimental results on three language pairs demonstrate that \method results in significant improvements over strong denoising auto-encoding baselines, with a gain of up to 1.3 BLEU and up to 9.2 entity accuracy points for English-Russian translation.