增强深层话题语义的对话引导模型

doi:10.3778/j.issn.1002-8331.2111-0527

摘要/Abstract

摘要： 目标导向式开放域对话的核心是根据对话历史与当前话题进行话题序列规划与回复检索，使得对话达到目标话题，话题特征的提取与建模方式影响话题预测的准确性，从而影响回复检索性能。目前常见的方法是引入外部知识来增强语义，但这种方法依赖外部知识质量。提出一种基于图卷积与子图注意力的话题语义增强的对话引导模型（KWGE），该模型首先构建关键词无向图，利用图卷积神经网络对关键词编码以增强话题语义，同时编码对话历史特征，使用对话特征与话题特征做预测与检索任务，每次预测任务后计算与当前话题相关话题子图的注意力权重来更新子图节点表示，用于获取更近于目标的话题。通过对两个真实闲聊数据集进行的广泛实验，表明该模型可以构建话题连贯的对话序列，并以较高的目标达成率实现对话目标。

关键词: 对话系统, 话题引导, 图卷积神经网络, 语义匹配

Abstract: The core of target-guided open-domain conversation is topic sequence planning and response retrieval according to the dialogue history and current topic, in that way the conversation reaches the target topic. The extraction and modeling method of topic features affects the accuracy of topic prediction, thereby affecting the performance of response retrieval. The current common method is to introduce external knowledge to enhance topic semantics, but this method relies on the quality of external knowledge. This paper proposes a topic semantics enhancement target-guided conversation model based on graph convolution and subgraph attention. The model firstly constructs an undirected graph of keywords, and uses a graph convolutional neural network to encode keywords to enhance topic semantics. At the same time, the dialogue history features are encoded, and the dialogue features and topic features are used for prediction and retrieval tasks. After each prediction task, the attention weight of the topic subgraph related to the current topic is calculated to update the sub-graph node representation, which is used to obtain information that is closer to the target topic. Extensive experiments on two real chat datasets show that the model can construct a coherent conversation sequence and achieve the conversation goals with a high achievement rate.

Key words: dialogue system, topic-guided, graph convolutional network, semantic matching

冯雅茹, 黄贤英, 李伟. 增强深层话题语义的对话引导模型[J]. 计算机工程与应用, 2023, 59(7): 171-179.

FENG Yaru, HUANG Xianying, LI Wei. Target-Guided Conversation Model Combining Enhanced Deep Topic Semantics[J]. Computer Engineering and Applications, 2023, 59(7): 171-179.

参考文献

[1] WEIZENBAUM J.ELIZA—a computer program for the study of natural language communication between man and machine[J].Communications of the ACM，1983，26（1）.
[2] VINYALS O，LE Q.A neural conversational model[J].Computer Science，2015.
[3] GHAZVININEJAD M，BROCKETT C，CHANG M W，et al.A knowledge-grounded neural conversation model[C]//Proceedings of the 32th AAAI Conference on Artificial Intelligence，New Orleans，Feb 2-7，2018：5110-5117.
[4] HIRANO T，HIGASHINAKA R，MATSUO Y.Analyzing post-dialogue comments by speakers-how do humans personalize their utterances in dialogue?[C]//Proceedings of the 17th Annual Meeting of the Special Interest Group on Discourse and Dialogue，Los Angeles，Sep 13-15，2016：157-165.
[5] TANG J，ZHAO T，XIONG C，et al.Target-guided open-domain conversation[C]//Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics，Florence，July 28-Aug 2，2019：5624-5634.
[6] SCHUSTER M，PALIWAL K K.Bidirectional recurrent neural networks[J].IEEE Transactions on Signal Processing，1997，45（11）：2673-2681.
[7] QIN J，YE Z，TANG J，et al.Dynamic knowledge routing network for target-guided open-domain conversation[C]//Proceedings of the 34th AAAI Conference on Artificial Intelligence，New York，Feb 7-12，2020：8657-8664.
[8] XU J，WANG H，NIU Z，et al.Knowledge graph grounded goal planning for open-domain conversation generation[C]//Proceedings of the 34th AAAI Conference on Artificial Intelligence，New York，Feb 7-12，2020：9338-9345.
[9] ZHONG P，LIU Y，WANG H，et al.Keyword-guided neural conversational model[C]//Proceedings of the 35th AAAI Conference on Artificial Intelligence，New York，Feb 2-9，2021：14568-14576.
[10] KIPF T N，WELLING M.Semi-supervised classification with graph convolutional networks[C]//Proceedings of the 5th International Conference on Learning Representations，Toulon，Apr 24-26，2017.
[11] 叶雪梅，毛雪岷，夏锦春，等.文本分类TF-IDF算法的改进研究[J].计算机工程与应用，2019，55（2）：104-109.
YE X M，MAO X M，XIA J C，et al.Improved approach to TF-IDF algorithm in text classification[J].Computer Engineering and Applications，2019，55（2）：104-109.
[12] WU Y，WU W，XING C，et al.Sequential matching network：a new architecture for multi-turn response selection in retrieval-based chatbots[C]//Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics，Vancouver，July 30-Aug 4，2017.Stroudsburg：ACL Press，2017：496-505.
[13] 刘欢.人机对话质量评估技术研究[D].西安：西安电子科技大学，2019.
LIU H.Research on quality estimation of human-machine interaction[D].Xi’an：Xidian University，2019.
[14] 王艳秋，管浩言，张彤.聊天机器人的分类标准和评估标准综述[J].软件工程，2021，24（2）：2-8.
WANG Y Q，GUAN H Y，ZHANG T.Research on taxonomy and evaluation criteria of chatbots[J].Software Engineering，2021，24（2）：2-8.
[15] 吴承泽.面向社区论坛多轮对话线索的连贯性评估方法研究[D].哈尔滨：哈尔滨工业大学，2018.
WU C Z.Research on coherence evaluation in the multi-turn dialogue threads base on community forums[D].Harbin：Harbin Industrial University，2018.
[16] 张伟男，张杨子，刘挺.对话系统评价方法综述[J].中国科学：信息科学，2017，47（8）：953-966.
ZHANG W N，ZHANG Y Z，LIU T.Survey of evaluation methods for dialogue system[J].Scientia Sinica Informationis，2017，47（8）：953-966.
[17] 叶铱雷，曹斌，范菁，等.面向任务型多轮对话的粗粒度意图识别方法[J].小型微型计算机系统，2020，41（8）：1620-1626.
YE Y L，CAO B，FAN J，et al.Coarse-grained intent recognition method for task-oriented multi-turn dialogue[J].Journal of Chinese Computer Systems，2020，41（8）：1620-1626.
[18] LI J，GALLEY M，BROCKETT C，et al.A persona-based neural conversation model[C]//Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics，Aug 7-12，2016.
[19] XU J，WANG H，NIU Z Y，et al.Conversational graph grounded policy learning for open-domain conversation generation[C]//Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics，July 5-10，2020.
[20] LEWIS M，YARATS D，DAUPHIN Y，et al.Deal or no deal? end-to-end learning of negotiation dialogues[C]//Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing，Copenhagen，Sep 9-11，2017：2443-2453
[21] 乔伟涛，黄海燕，王珊.基于Transformer编码器的语义相似度算法研究[J].计算机工程与应用，2021，57（14）：158-163.
QIAO W T，HUANG H Y，WANG S.Semantic similarity calculation based on Transformer encoder[J].Computer Engineering and Applications，2021，57（14）：158-163.
[22] 卢琪，潘志松，谢钧.融合知识表示学习的双向注意力问答模型[J].计算机工程与应用，2021，57（23）：171-177.
LU Q，PAN Z S，XIE J.Bidirectional attention question answering model combining knowledge representation learning[J].Computer Engineering and Applications，2021，57（23）：171-177.
[23] 文敦伟，樊小虎.一个基于语义模块的交互式本体匹配框架[J].计算机工程与应用，2007，43（29）：63-67.
WEN D W，FAN X H.Semi-auto framework for module based ontology matching[J].Computer Engineering and Applications，2007，43（29）：63-67.