Abstract
Neural machine translation (NMT) takes deterministic sequences for source representations. However, either word-level or subword-level segmentations have multiple choices to split a source sequence with different word segmentors or different subword vocabulary sizes. We hypothesize that the diversity in segmentations may affect the NMT performance. To integrate different segmentations with the state-of-the-art NMT model, Transformer, we propose lattice-based encoders to explore effective word or subword representation in an automatic way during training. We propose two methods: 1) lattice positional encoding and 2) lattice-aware self-attention. These two methods can be used together and show complementary to each other to further improve translation performance. Experiment results show superiorities of lattice-based encoders in word-level and subword-level representations over conventional Transformer encoder.- Anthology ID:
- P19-1298
- Volume:
- Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics
- Month:
- July
- Year:
- 2019
- Address:
- Florence, Italy
- Venue:
- ACL
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 3090–3097
- Language:
- URL:
- https://aclanthology.org/P19-1298
- DOI:
- 10.18653/v1/P19-1298
- Cite (ACL):
- Fengshun Xiao, Jiangtong Li, Hai Zhao, Rui Wang, and Kehai Chen. 2019. Lattice-Based Transformer Encoder for Neural Machine Translation. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 3090–3097, Florence, Italy. Association for Computational Linguistics.
- Cite (Informal):
- Lattice-Based Transformer Encoder for Neural Machine Translation (Xiao et al., ACL 2019)
- PDF:
- https://preview.aclanthology.org/starsem-semeval-split/P19-1298.pdf