Abstract
Real world deployments of word alignment are almost certain to cover both high and low resource languages. However, the state-of-the-art for this task recommends a different model class depending on the availability of gold alignment training data for a particular language pair. We propose BinaryAlign, a novel word alignment technique based on binary sequence labeling that outperforms existing approaches in both scenarios, offering a unifying approach to the task. Additionally, we vary the specific choice of multilingual foundation model, perform stratified error analysis over alignment error type, and explore the performance of BinaryAlign on non-English language pairs. We make our source code publicly available.- Anthology ID:
- 2024.acl-long.553
- Volume:
- Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
- Month:
- August
- Year:
- 2024
- Address:
- Bangkok, Thailand
- Editors:
- Lun-Wei Ku, Andre Martins, Vivek Srikumar
- Venue:
- ACL
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 10277–10288
- Language:
- URL:
- https://aclanthology.org/2024.acl-long.553
- DOI:
- 10.18653/v1/2024.acl-long.553
- Cite (ACL):
- Gaetan Latouche, Marc-André Carbonneau, and Benjamin Swanson. 2024. BinaryAlign: Word Alignment as Binary Sequence Labeling. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 10277–10288, Bangkok, Thailand. Association for Computational Linguistics.
- Cite (Informal):
- BinaryAlign: Word Alignment as Binary Sequence Labeling (Latouche et al., ACL 2024)
- PDF:
- https://preview.aclanthology.org/autopr/2024.acl-long.553.pdf