Improving Lexically Constrained Neural Machine Translation with Source-Conditioned Masked Span Prediction
This work addresses the problem of accurate terminology translation for domain-specific applications, representing an incremental improvement over existing methods.
The paper tackled the challenge of translating domain-specific corpora with long n-grams and specialized terms in lexically constrained neural machine translation, proposing a training strategy that improved terminology and sentence-level translation across three domain-specific corpora in two language pairs.
Accurate terminology translation is crucial for ensuring the practicality and reliability of neural machine translation (NMT) systems. To address this, lexically constrained NMT explores various methods to ensure pre-specified words and phrases appear in the translation output. However, in many cases, those methods are studied on general domain corpora, where the terms are mostly uni- and bi-grams (>98%). In this paper, we instead tackle a more challenging setup consisting of domain-specific corpora with much longer n-gram and highly specialized terms. Inspired by the recent success of masked span prediction models, we propose a simple and effective training strategy that achieves consistent improvements on both terminology and sentence-level translation for three domain-specific corpora in two language pairs.