Domain Robustness in Neural Machine Translation

TL;DR

This paper investigates the domain robustness of SMT and NMT models, highlighting distinct issues like fluency and adequacy, and evaluates methods to improve their performance on out-of-domain translation tasks.

Contribution

It provides a comparative analysis of SMT and NMT domain robustness issues and empirically evaluates methods to enhance out-of-domain translation quality.

Findings

NMT suffers from hallucinations causing low adequacy in unknown domains.

Several methods improve domain robustness and BLEU scores.

Methods only slightly increase translation adequacy despite higher BLEU scores.

Abstract

Translating text that diverges from the training domain is a key challenge for machine translation. Domain robustness---the generalization of models to unseen test domains---is low for both statistical (SMT) and neural machine translation (NMT). In this paper, we study the performance of SMT and NMT models on out-of-domain test sets. We find that in unknown domains, SMT and NMT suffer from very different problems: SMT systems are mostly adequate but not fluent, while NMT systems are mostly fluent, but not adequate. For NMT, we identify such hallucinations (translations that are fluent but unrelated to the source) as a key reason for low domain robustness. To mitigate this problem, we empirically compare methods that are reported to improve adequacy or in-domain robustness in terms of their effectiveness at improving domain robustness. In experiments on German to English OPUS data, and German to Romansh (a low-resource setting) we find that several methods improve domain robustness. While those methods do lead to higher BLEU scores overall, they only slightly increase the adequacy of translations compared to SMT.