Abstract
Neural machine translation models have little inductive bias, which can be a disadvantage in low-resource scenarios. They require large volumes of data and often perform poorly when limited data is available. We show that using naive regularization methods, based on sentence length, punctuation and word frequencies, to penalize translations that are very different from the input sentences, consistently improves the translation quality across multiple low-resource languages. We experiment with 12 language pairs, varying the training data size between 17k to 230k sentence pairs. Our best regularizer achieves an average increase of 1.5 BLEU score and 1.0 TER score across all the language pairs. For example, we achieve a BLEU score of 26.70 on the IWSLT15 English-Vietnamese translation task simply by using relative differences in punctuation as a regularizer.
Original language | English |
---|---|
Title of host publication | International Conference on Recent Advances in Natural Language Processing in a Deep Learning World, RANLP 2019 - Proceedings |
Editors | Galia Angelova, Ruslan Mitkov, Ivelina Nikolova, Irina Temnikova, Irina Temnikova |
Number of pages | 10 |
Publisher | Incoma Ltd |
Publication date | 2019 |
Pages | 102-111 |
ISBN (Electronic) | 9789544520557 |
DOIs | |
Publication status | Published - 2019 |
Event | 12th International Conference on Recent Advances in Natural Language Processing, RANLP 2019 - Varna, Bulgaria Duration: 2 Sep 2019 → 4 Sep 2019 |
Conference
Conference | 12th International Conference on Recent Advances in Natural Language Processing, RANLP 2019 |
---|---|
Country/Territory | Bulgaria |
City | Varna |
Period | 02/09/2019 → 04/09/2019 |