Abstract
Contextualized representations give significantly improved results for a wide range of NLP tasks. Much work has been dedicated to analyzing the features captured by representative models such as BERT. Existing work finds that syntactic, semantic and word sense knowledge are encoded in BERT. However, little work has investigated word features for character languages such as Chinese. We investigate Chinese BERT using both attention weight distribution statistics and probing tasks, finding that (1) word information is captured by BERT; (2) word-level features are mostly in the middle representation layers; (3) downstream tasks make different use of word features in BERT, with POS tagging and chunking relying the most on word features, and natural language inference relying the least on such features.- Anthology ID:
- 2020.coling-main.254
- Volume:
- Proceedings of the 28th International Conference on Computational Linguistics
- Month:
- December
- Year:
- 2020
- Address:
- Barcelona, Spain (Online)
- Editors:
- Donia Scott, Nuria Bel, Chengqing Zong
- Venue:
- COLING
- SIG:
- Publisher:
- International Committee on Computational Linguistics
- Note:
- Pages:
- 2826–2836
- Language:
- URL:
- https://aclanthology.org/2020.coling-main.254
- DOI:
- 10.18653/v1/2020.coling-main.254
- Cite (ACL):
- Yile Wang, Leyang Cui, and Yue Zhang. 2020. Does Chinese BERT Encode Word Structure?. In Proceedings of the 28th International Conference on Computational Linguistics, pages 2826–2836, Barcelona, Spain (Online). International Committee on Computational Linguistics.
- Cite (Informal):
- Does Chinese BERT Encode Word Structure? (Wang et al., COLING 2020)
- PDF:
- https://preview.aclanthology.org/nschneid-patch-4/2020.coling-main.254.pdf
- Code
- ylwangy/BERT_zh_Analysis
- Data
- CLUE, CMNLI, WSC