Subword Segmental Machine Translation: Unifying Segmentation and Target Sentence Generation
Authors:
Preprint
Abstract
Subword segmenters like BPE operate as a preprocessing step in neural machine translation and other (conditional) language models. They are applied to datasets before training, so translation or text generation quality relies on the quality of segmentations. We propose a departure from this paradigm, called subword segmental machine translation (SSMT). SSMT unifies subword segmentation and MT in a single trainable model. It learns to segment target sentence words while jointly learning to generate target sentences. To use SSMT during inference we propose dynamic decoding, a text generation algorithm that adapts segmentations as it generates translations. Experiments across 6 translation directions show that SSMT improves chrF scores for morphologically rich agglutinative languages. Gains are strongest in the very low-resource scenario. SSMT also learns subwords that are closer to morphemes compared to baselines and proves more robust on a test set constructed for evaluating morphological compositional generalisation.Keywords
All Related Versions
This publication has 21 references indexed in Scilit:
- Linguistically annotated dataset for four official South African languages with a conjunctive orthography: IsiNdebele, isiXhosa, isiZulu, and SiswatiData in Brief, 2022
- Canine: Pre-training an Efficient Tokenization-Free Encoder for Language RepresentationTransactions of the Association for Computational Linguistics, 2022
- Why don’t people use character-level machine translation?2022
- Multi-view Subword Regularization2021
- Optimizing Transformer for Low-Resource Neural Machine Translation2020
- COGS: A Compositional Generalization Challenge Based on Semantic Interpretation2020
- Getting the ##life out of living: How Adequate Are Word-Pieces for Modelling Complex Morphology?2020
- Dynamic Programming Encoding for Subword Segmentation in Neural Machine Translation2020
- Measures of distance between probability distributionsJournal of Mathematical Analysis and Applications, 1989
- Connectionism and cognitive architecture: A critical analysisCognition, 1988