Ti-Reader: 基于注意力机制的藏文机器阅读理解端到端网络模型(Ti-Reader: An End-to-End Network Model Based on Attention Mechanisms for Tibetan Machine Reading Comprehension)

Yuan Sun (孙媛), Chaofan Chen (陈超凡), Sisi Liu (刘思思), Xiaobing Zhao (赵小兵)


Abstract
机器阅读理解旨在教会机器去理解一篇文章并且回答与之相关的问题。为了解决低资源语言上机器阅读理解模型性能低的问题,本文提出了一种基于注意力机制的藏文机器阅读理解端到端网络模型Ti-Reader。首先,为了编码更细粒度的藏文文本信息,本文将音节和词相结合进行词表示,然后采用词级注意力机制去关注文本中的关键词,采用重读机制去捕捉文章和问题之间的语义信息,采用自注意力机制去匹配问题与答案的隐变量本身,为答案预测提供更多的线索。最后,实验结果表明,Ti-Reader模型提升了藏文机器阅读理解的性能,并且在英文数据集SQuAD上也有较好的表现。
Anthology ID:
2021.ccl-1.21
Volume:
Proceedings of the 20th Chinese National Conference on Computational Linguistics
Month:
August
Year:
2021
Address:
Huhhot, China
Venue:
CCL
SIG:
Publisher:
Chinese Information Processing Society of China
Note:
Pages:
219–228
Language:
Chinese
URL:
https://aclanthology.org/2021.ccl-1.21
DOI:
Bibkey:
Cite (ACL):
Yuan Sun, Chaofan Chen, Sisi Liu, and Xiaobing Zhao. 2021. Ti-Reader: 基于注意力机制的藏文机器阅读理解端到端网络模型(Ti-Reader: An End-to-End Network Model Based on Attention Mechanisms for Tibetan Machine Reading Comprehension). In Proceedings of the 20th Chinese National Conference on Computational Linguistics, pages 219–228, Huhhot, China. Chinese Information Processing Society of China.
Cite (Informal):
Ti-Reader: 基于注意力机制的藏文机器阅读理解端到端网络模型(Ti-Reader: An End-to-End Network Model Based on Attention Mechanisms for Tibetan Machine Reading Comprehension) (Sun et al., CCL 2021)
Copy Citation:
PDF:
https://preview.aclanthology.org/ingestion-script-update/2021.ccl-1.21.pdf
Data
SQuAD