基于BERT与柱搜索的中文释义生成(Chinese Definition Modeling Based on BERT and Beam Seach)
Qinan Fan (范齐楠), Cunliang Kong (孔存良), Liner Yang (杨麟儿), Erhong Yang (杨尔弘)
Abstract
释义生成任务是指为一个目标词生成相应的释义。前人研究中文释义生成任务时未考虑目标词的上下文,本文首次在中文释义生成任务中使用了目标词的上下文信息,并提出了一个基于BERT与柱搜索的释义生成模型。本文构建了包含上下文的CWN中文数据集用于开展实验,除了BLEU指标之外,还使用语义相似度作为额外的自动评价指标,实验结果显示本文模型在中文CWN数据集和英文Oxford数据集上均有显著提升,人工评价结果也与自动评价结果一致。最后,本文对生成实例进行了深入分析。- Anthology ID:
- 2020.ccl-1.32
- Volume:
- Proceedings of the 19th Chinese National Conference on Computational Linguistics
- Month:
- October
- Year:
- 2020
- Address:
- Haikou, China
- Venue:
- CCL
- SIG:
- Publisher:
- Chinese Information Processing Society of China
- Note:
- Pages:
- 336–348
- Language:
- Chinese
- URL:
- https://aclanthology.org/2020.ccl-1.32
- DOI:
- Cite (ACL):
- Qinan Fan, Cunliang Kong, Liner Yang, and Erhong Yang. 2020. 基于BERT与柱搜索的中文释义生成(Chinese Definition Modeling Based on BERT and Beam Seach). In Proceedings of the 19th Chinese National Conference on Computational Linguistics, pages 336–348, Haikou, China. Chinese Information Processing Society of China.
- Cite (Informal):
- 基于BERT与柱搜索的中文释义生成(Chinese Definition Modeling Based on BERT and Beam Seach) (Fan et al., CCL 2020)
- PDF:
- https://preview.aclanthology.org/ingestion-script-update/2020.ccl-1.32.pdf