Abstract
Exposing diverse subword segmentations to neural machine translation (NMT) models often improves the robustness of machine translation as NMT models can experience various subword candidates. However, the diversification of subword segmentations mostly relies on the pre-trained subword language models from which erroneous segmentations of unseen words are less likely to be sampled. In this paper, we present adversarial subword regularization (ADVSR) to study whether gradient signals during training can be a substitute criterion for exposing diverse subword segmentations. We experimentally show that our model-based adversarial samples effectively encourage NMT models to be less sensitive to segmentation errors and improve the performance of NMT models in low-resource and out-domain datasets.- Anthology ID:
- 2020.findings-emnlp.175
- Volume:
- Findings of the Association for Computational Linguistics: EMNLP 2020
- Month:
- November
- Year:
- 2020
- Address:
- Online
- Editors:
- Trevor Cohn, Yulan He, Yang Liu
- Venue:
- Findings
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 1945–1953
- Language:
- URL:
- https://aclanthology.org/2020.findings-emnlp.175
- DOI:
- 10.18653/v1/2020.findings-emnlp.175
- Cite (ACL):
- Jungsoo Park, Mujeen Sung, Jinhyuk Lee, and Jaewoo Kang. 2020. Adversarial Subword Regularization for Robust Neural Machine Translation. In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 1945–1953, Online. Association for Computational Linguistics.
- Cite (Informal):
- Adversarial Subword Regularization for Robust Neural Machine Translation (Park et al., Findings 2020)
- PDF:
- https://preview.aclanthology.org/naacl24-info/2020.findings-emnlp.175.pdf
- Code
- dmis-lab/AdvSR
- Data
- MTNT