CET2: Modelling Topic Transitions for Coherent and Engaging Knowledge-Grounded Conversations

被引:0
|
作者
Xu, Lin [1 ,2 ]
Zhou, Qixian [3 ]
Fu, Jinlan [2 ]
Ng, See-Kiong [1 ,2 ]
机构
[1] Natl Univ Singapore, Sch Comp, Dept Comp Sci, Singapore 117417, Singapore
[2] Natl Univ Singapore, Inst Data Sci, Singapore 117602, Singapore
[3] Bytedance, Shenzhen 518066, Peoples R China
基金
新加坡国家研究基金会;
关键词
Oral communication; Coherence; Task analysis; Dogs; Gold; Training; Image color analysis; Dialogue topic transition; knowledge grounded dialogue system; knowledge selection;
D O I
10.1109/TASLP.2023.3313418
中图分类号
O42 [声学];
学科分类号
070206 ; 082403 ;
摘要
Knowledge-grounded dialogue systems aim to generate coherent and engaging responses based on the dialogue contexts and selected external knowledge. Previous knowledge selection methods tend to rely too heavily on the dialogue contexts or over-emphasize the new information in the selected knowledge, resulting in the selection of repetitious or incongruous knowledge and further generating repetitive or incoherent responses, as the generation of the response depends on the chosen knowledge. To address these shortcomings, we introduce a Coherent and Engaging Topic Transition (CET2) framework to model topic transitions for selecting knowledge that is coherent to the context of the conversations while providing adequate knowledge diversity for topic development. Our CET2 framework considers multiple factors for knowledge selection, including valid transition logic from dialogue contexts to the following topics and systematic comparisons between available knowledge candidates. Extensive experiments on two public benchmarks demonstrate the superiority and the better generalization ability of CET2 on knowledge selection. This is due to our well-designed transition features and comparative knowledge selection strategy, which are more transferable to conversations about unseen topics. Analysis of fine-grained knowledge selection accuracy also shows that CET2 can better balance topic entailment (contextual coherence) and development (knowledge diversity) in dialogue than existing approaches.
引用
收藏
页码:3527 / 3536
页数:10
相关论文
共 12 条
  • [1] Augmenting Knowledge-grounded Conversations with Sequential Knowledge Transition
    Zhan, Haolan
    Zhang, Hainan
    Chen, Hongshen
    Ding, Zhuoye
    Bao, Yongjun
    Lan, Yanyan
    [J]. 2021 CONFERENCE OF THE NORTH AMERICAN CHAPTER OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS: HUMAN LANGUAGE TECHNOLOGIES (NAACL-HLT 2021), 2021, : 5621 - 5630
  • [2] Knowledge-Grounded Dialogue Generation for Medical Conversations: A Survey
    Liu, Xiaoxiao
    Chang, Jian
    Zhang, Jian Jun
    [J]. 2023 27TH INTERNATIONAL CONFERENCE INFORMATION VISUALISATION, IV, 2023, : 409 - 413
  • [3] TopicKS: Topic-driven Knowledge Selection for Knowledge-grounded Dialogue Generation
    Wang, Shiquan
    Si, Yuke
    Wei, Xiao
    Wang, Longbiao
    Zhuang, Zhiqiang
    Zhang, Xiaowang
    Dang, Jianwu
    [J]. INTERSPEECH 2022, 2022, : 1121 - 1125
  • [4] Initiative-Aware Self-Supervised Learning for Knowledge-Grounded Conversations
    Meng, Chuan
    Ren, Pengjie
    Chen, Zhumin
    Ren, Zhaochun
    Xi, Tengxiao
    de Rijke, Maarten
    [J]. SIGIR '21 - PROCEEDINGS OF THE 44TH INTERNATIONAL ACM SIGIR CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION RETRIEVAL, 2021, : 522 - 532
  • [5] Topical-Chat: Towards Knowledge-Grounded Open-Domain Conversations
    Gopalakrishnan, Karthik
    Hedayatnia, Behnam
    Chen, Qinlang
    Gottardi, Anna
    Kwatra, Sanjeev
    Venkatesh, Anu
    Gabriel, Raefer
    Hakkani-Tur, Dilek
    [J]. INTERSPEECH 2019, 2019, : 1891 - 1895
  • [6] Is a Knowledge-based Response Engaging?: An Analysis on Knowledge-Grounded Dialogue with Information Source Annotation
    Kodama, Takashi
    Kiyomaru, Hirokazu
    Huang, Yin Jou
    Okahisa, Taro
    Kurohashi, Sadao
    [J]. PROCEEDINGS OF THE 61ST ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, ACL-SRW 2023, VOL 4, 2023, : 237 - 243
  • [7] A Pre-training Strategy for Zero-Resource Response Selection in Knowledge-Grounded Conversations
    Tao, Chongyang
    Chen, Changyu
    Feng, Jiazhan
    Wen, Ji-Rong
    Yan, Rui
    [J]. 59TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS AND THE 11TH INTERNATIONAL JOINT CONFERENCE ON NATURAL LANGUAGE PROCESSING (ACL-IJCNLP 2021), VOL 1, 2021, : 4446 - 4457
  • [8] Building knowledge-grounded dialogue systems with graph-based semantic modelling
    Yang, Yizhe
    Huang, Heyan
    Gao, Yang
    Li, Jiawei
    [J]. KNOWLEDGE-BASED SYSTEMS, 2024, 298
  • [9] Enhancing Knowledge Selection with Data Processing Based on Multiple Turns of Dialog in Knowledge-Grounded Open-Domain Conversations
    Joo, Eojin
    Lee, Do Kyung
    Youn, Junyoung
    Sung, Joo-Won
    Choi, Ho-Jin
    [J]. 2024 IEEE INTERNATIONAL CONFERENCE ON BIG DATA AND SMART COMPUTING, IEEE BIGCOMP 2024, 2024, : 365 - 366
  • [10] Knowledge-grounded dialogue modelling with dialogue-state tracking, domain tracking, and entity extraction
    Hong, Taesuk
    Cho, Junhee
    Yu, Haeun
    Ko, Youngjoong
    Seo, Jungyun
    [J]. COMPUTER SPEECH AND LANGUAGE, 2023, 78