Abstract
The span-based model enjoys great popularity in recent works of sequence segmentation. However, each of these methods suffers from its own defects, such as invalid predictions. In this work, we introduce a unified span-based model, lexical unit analysis (LUA), that addresses all these matters. Segmenting a lexical unit sequence involves two steps. Firstly, we embed every span by using the representations from a pretraining language model. Secondly, we define a score for every segmentation candidate and apply dynamic programming (DP) to extract the candidate with the maximum score. We have conducted extensive experiments on 3 tasks, (e.g., syntactic chunking), across 7 datasets. LUA has established new state-of-the-art performances on 6 of them. We have achieved even better results through incorporating label correlations.- Anthology ID:
- 2021.findings-emnlp.18
- Volume:
- Findings of the Association for Computational Linguistics: EMNLP 2021
- Month:
- November
- Year:
- 2021
- Address:
- Punta Cana, Dominican Republic
- Venue:
- Findings
- SIG:
- SIGDAT
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 181–187
- Language:
- URL:
- https://aclanthology.org/2021.findings-emnlp.18
- DOI:
- 10.18653/v1/2021.findings-emnlp.18
- Cite (ACL):
- Yangming Li, Lemao Liu, and Shuming Shi. 2021. Segmenting Natural Language Sentences via Lexical Unit Analysis. In Findings of the Association for Computational Linguistics: EMNLP 2021, pages 181–187, Punta Cana, Dominican Republic. Association for Computational Linguistics.
- Cite (Informal):
- Segmenting Natural Language Sentences via Lexical Unit Analysis (Li et al., Findings 2021)
- PDF:
- https://preview.aclanthology.org/ingestion-script-update/2021.findings-emnlp.18.pdf
- Data
- ATIS, SNIPS, Universal Dependencies