Paying Attention to Multi-Word Expressions in Neural Machine Translation
This addresses translation accuracy for languages with complex expressions, but it is incremental as it builds on existing NMT methods with data augmentation.
The paper tackled the problem of neural machine translation struggling with multi-word expressions by adding automatically extracted MWE candidates to training data, resulting in a best improvement of 0.99 BLEU points for English->Latvian and English->Czech systems.
Processing of multi-word expressions (MWEs) is a known problem for any natural language processing task. Even neural machine translation (NMT) struggles to overcome it. This paper presents results of experiments on investigating NMT attention allocation to the MWEs and improving automated translation of sentences that contain MWEs in English->Latvian and English->Czech NMT systems. Two improvement strategies were explored -(1) bilingual pairs of automatically extracted MWE candidates were added to the parallel corpus used to train the NMT system, and (2) full sentences containing the automatically extracted MWE candidates were added to the parallel corpus. Both approaches allowed to increase automated evaluation results. The best result - 0.99 BLEU point increase - has been reached with the first approach, while with the second approach minimal improvements achieved. We also provide open-source software and tools used for MWE extraction and alignment inspection.