|本期目录/Table of Contents|

[1]陈巧红,李妃玉,贾宇波,等.基于自注意力和门控机制的答案选择模型[J].浙江理工大学学报,2021,45-46(自科三):400-407.
　CHEN Qiaohong,LI Feiyu,JIA Yubo,et al.Answer selection model based on selfattention and gating mechanism[J].Journal of Zhejiang Sci-Tech University,2021,45-46(自科三):400-407.
点击复制

基于自注意力和门控机制的答案选择模型()
分享到：

浙江理工大学学报[ISSN:1673-3851/CN:33-1338/TS]

卷:
第45-46卷
期数:
2021年自科第三期
页码:
400-407
栏目:
出版日期:
2021-04-28

文章信息/Info

Title:
Answer selection model based on selfattention and gating mechanism
文章编号:
1673-3851 (2021) 05-0400-08
作者:
陈巧红; 李妃玉; 贾宇波; 孙麒
浙江理工大学信息学院，杭州 310018
Author(s):
CHEN Qiaohong;  LI Feiyu;  JIA Yubo;  SUN Qi
School of Information Science and Technology, Zhejiang Sci-Tech University, Hangzhou 310018, China
关键词:
答案选择; 层叠自注意力; 注意力机制; 门控机制
分类号:
TP391
文献标志码:
A
摘要:
针对现有答案选择方法语义特征提取不充分和准确性差的问题，引入自注意力和门控机制，提出了一种答案选择模型。该模型首先在问题和答案文本内部利用层叠自注意力进行向量表示，并在自注意力模块中让单词和位置分开进行多头注意力；然后将答案句通过卷积神经网络（Convolutional neural network, CNN）得到的向量表示输入注意力层，根据问题生成与问题相关的答案表示，并通过门控机制融合两种表示；最后计算问题和答案文本的相关性分数，得到候选答案的排名和标注。结果表明：该模型与双向长短时记忆网络模型、自注意力模型和基于注意力的双向长短时记忆网络模型相比，在WebMedQA数据集上平均倒数排名分数分别提高了837%、479%和203%，预测答案正确率也有提高。这表明提出的模型能够捕获更丰富的语义信息，有效提升了答案选择的性能。

参考文献/References:

［1］ Fan H J, Ma Z Y, Li H Q, et al. Enhanced answer selection in CQA using multidimensional features combination［J］. Tsinghua Science and Technology, 2019, 24(3): 346-359.
［2］ Heilman M, Smith N A. Tree edit models for recognizing textual entailments, paraphrases, and answers to questions［C］//Human Language Technologies: Conference of the North American Chapter of the Association of Computational Linguistics. California, USA: Association for Computing Machinery, 2010: 1011-1019.
［3］ Chan W, Zhou X D, Wang W, et al. Community answer summarization for multisentence question with group L1 regularization ［C］//Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics. Jeju Island, Korea: Association for Computational Linguistics, 2012: 582-591.
［4］ Chen Z W, He Z, Liu X W, et al. Evaluating semantic relations in neural word embeddings with biomedical and general domain knowledge bases［J］. BMC Medical Informatics & Decision Making, 2018, 18(2):53-68.
［5］ Zhou X Q, Hu B T, Chen Q C, et al. Answer sequence learning with neural networks for answer selection in community question answering ［C］//Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics. Beijing, China: Association for Computational Linguistics, 2015: 713-718.
［6］ Wakchaure M, Kulkarni P. A scheme of answer selection in community question answering using machine learning techniques ［C］//2019 International Conference on Intelligent Computing and Control Systems (ICCS). Madurai, India: IEEE, 2019: 879-883.
［7］ Tan M, dos Santos C, Xiang B, et al. Improved representation learning for question answer matching［C］//Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics. Berlin, Germany: Association for Computational Linguistics, 2016: 464-473.
［8］ 熊雪, 刘秉权, 吴翔虎. 基于注意力机制的答案选择方法研究［J］. 智能计算机与应用, 2018, 8(6):90-93.
［9］ Shao T H, Guo Y P, Chen H H, et al. Transformerbased neural network for answer selection in question answering［J］. IEEE Access, 2019, 99(7):146-156.
［10］ 谢正文, 柏钧献, 熊熙, 等. 基于增强问题重要性表示的答案选择算法研究［J］. 四川大学学报(自然科学版), 2020, 57(1): 66-72.

备注/Memo

备注/Memo:
收稿日期：2021-02-01
网络出版日期：2021-03-30
基金项目：浙江理工大学中青年骨干人才培养经费项目
作者简介：陈巧红(1978-)，女，浙江临海人，副教授，博士，主要从事计算机辅助设计及机器学习技术方面的研究
更新日期/Last Update: 2021-06-29