Learning and Evaluating Chinese Idiom Embeddings

Minghuan Tan, Jing Jiang


Abstract
We study the task of learning and evaluating Chinese idiom embeddings. We first construct a new evaluation dataset that contains idiom synonyms and antonyms. Observing that existing Chinese word embedding methods may not be suitable for learning idiom embeddings, we further present a BERT-based method that directly learns embedding vectors for individual idioms. We empirically compare representative existing methods and our method. We find that our method substantially outperforms existing methods on the evaluation dataset we have constructed.
Anthology ID:
2021.ranlp-1.155
Volume:
Proceedings of the International Conference on Recent Advances in Natural Language Processing (RANLP 2021)
Month:
September
Year:
2021
Address:
Held Online
Editors:
Ruslan Mitkov, Galia Angelova
Venue:
RANLP
SIG:
Publisher:
INCOMA Ltd.
Note:
Pages:
1387–1396
Language:
URL:
https://aclanthology.org/2021.ranlp-1.155
DOI:
Bibkey:
Cite (ACL):
Minghuan Tan and Jing Jiang. 2021. Learning and Evaluating Chinese Idiom Embeddings. In Proceedings of the International Conference on Recent Advances in Natural Language Processing (RANLP 2021), pages 1387–1396, Held Online. INCOMA Ltd..
Cite (Informal):
Learning and Evaluating Chinese Idiom Embeddings (Tan & Jiang, RANLP 2021)
Copy Citation:
PDF:
https://preview.aclanthology.org/nschneid-patch-2/2021.ranlp-1.155.pdf
Code
 VisualJoyce/ChengyuBERT
Data
COS960ChID