MixEI: Mixing explicit and implicit commonsense knowledge in open-domain dialogue response generation

被引:0
|
作者
Wu, Sixing [1 ,2 ]
Yu, Jiong [1 ,2 ]
Zhou, Wei [1 ,2 ]
机构
[1] National Pilot School of Software, Yunnan University, Yunnan, Kunming,650504, China
[2] Engineering Research Center of Cyberspace, Yunnan University, Yunnan, Kunming,650504, China
基金
中国国家自然科学基金;
关键词
Knowledge transfer;
D O I
10.1016/j.neucom.2024.128999
中图分类号
学科分类号
摘要
The inadequate awareness of real-world knowledge often causes machines to produce generic responses, such as ‘I think so.’, which may bring the degression of user interests. Consequently, enriching knowledge awareness and fabricating informative responses are long-standing challenges in open-domain dialogue systems. Previous studies have shown incorporating everyday commonsense knowledge can significantly enhance the open-domain dialogue response models. Nonetheless, previous works can only use explicit or implicit knowledge. Unlike them, this work presents a novel MixEI to leverage both explicit and implicit commonsense knowledge in dialogue generation. MixEI uses Dual-Way Knowledge Alignment and MixEI-Ranker to retrieve a set of contextually relevant commonsense facts as the explicit background knowledge and identify implicit knowledge labels by selecting clue facts that can tightly connect the dialogue context. MixEI uses BART as the backbone. After jointly encoding the background knowledge and dialogue history, MixEI first tries to externalize the implicit clue knowledge; then, the response decoding can seek information from both explicit and implicit knowledge. Extensive experiments on Chinese Weibo and English Reddit have verified the superior performance of the proposed MixEI-Ranker and MixEI. © 2024 Elsevier B.V.
引用
收藏
相关论文
共 50 条
  • [31] Profile Consistency Identification for Open-domain Dialogue Agents
    Song, Haoyu
    Yan Wang
    Zhang, Wei Nan
    Zhao, Zhengyu
    Ting Liu
    Xiaojiang Liu
    [J]. PROCEEDINGS OF THE 2020 CONFERENCE ON EMPIRICAL METHODS IN NATURAL LANGUAGE PROCESSING (EMNLP), 2020, : 6651 - 6662
  • [32] Dialogue Distillation: Open-Domain Dialogue Augmentation Using Unpaired Data
    Zhang, Rongsheng
    Zheng, Yinhe
    Shao, Jianzhi
    Mao, Xiaoxi
    Xi, Yadong
    Huang, Minlie
    [J]. PROCEEDINGS OF THE 2020 CONFERENCE ON EMPIRICAL METHODS IN NATURAL LANGUAGE PROCESSING (EMNLP), 2020, : 3449 - 3460
  • [33] Unsupervised Open-domain Keyphrase Generation
    Lam Thanh Do
    Akash, Pritom Saha
    Chang, Kevin Chen-Chuan
    [J]. PROCEEDINGS OF THE 61ST ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2023): LONG PAPERS, VOL 1, 2023, : 10614 - 10627
  • [34] MODE: a multimodal open-domain dialogue dataset with explanation
    Yin, Hang
    Lu, Pinren
    Li, Ziang
    Sun, Bin
    Li, Kan
    [J]. APPLIED INTELLIGENCE, 2024, 54 (07) : 5891 - 5906
  • [35] Estimating User Interest from Open-Domain Dialogue
    Inaba, Michimasa
    Takahash, Kenichi
    [J]. 19TH ANNUAL MEETING OF THE SPECIAL INTEREST GROUP ON DISCOURSE AND DIALOGUE (SIGDIAL 2018), 2018, : 32 - 40
  • [36] SHONGLAP: A Large Bengali Open-Domain Dialogue Corpus
    Monsur, Syed Mostofa
    Chowdhury, Sakib
    Fatemi, Md Shahrar
    Ahmed, Shafayat
    [J]. LREC 2022: THIRTEEN INTERNATIONAL CONFERENCE ON LANGUAGE RESOURCES AND EVALUATION, 2022, : 5797 - 5804
  • [37] Enhancing the Open-Domain Dialogue Evaluation in Latent Space
    Chan, Zhangming
    Liu, Lemao
    Li, Juntao
    Zhang, Haisong
    Zhao, Dongyan
    Shi, Shuming
    Yan, Rui
    [J]. FINDINGS OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, ACL-IJCNLP 2021, 2021, : 4889 - 4900
  • [38] Towards Multilingual Automatic Open-Domain Dialogue Evaluation
    Mendonca, John
    Lavie, Alon
    Trancoso, Isabel
    [J]. 24TH MEETING OF THE SPECIAL INTEREST GROUP ON DISCOURSE AND DIALOGUE, SIGDIAL 2023, 2023, : 130 - 141
  • [39] Probing Commonsense Explanation in Dialogue Response Generation
    Zhou, Pei
    Jandaghi, Pegah
    Cho, Hyundong
    Lin, Bill Yuchen
    Pujara, Jay
    Ren, Xiang
    [J]. FINDINGS OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, EMNLP 2021, 2021, : 4132 - 4146
  • [40] Implicit Discourse Relation Identification for Open-domain Dialogues
    Ma, Mingyu Derek
    Bowden, Kevin K.
    Wu, Jiaqi
    Cui, Wen
    Walker, Marilyn
    [J]. 57TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2019), 2019, : 666 - 672