@inproceedings{li-etal-2024-helpful,
    title = "Be Helpful but Don`t Talk too Much - Enhancing Helpfulness in Conversations through Relevance in Multi-Turn Emotional Support",
    author = "Li, Junlin  and
      Peng, Bo  and
      Hsu, Yu-Yin  and
      Huang, Chu-Ren",
    editor = "Al-Onaizan, Yaser  and
      Bansal, Mohit  and
      Chen, Yun-Nung",
    booktitle = "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing",
    month = nov,
    year = "2024",
    address = "Miami, Florida, USA",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/jlcl-multiple-ingestion/2024.emnlp-main.118/",
    doi = "10.18653/v1/2024.emnlp-main.118",
    pages = "1976--1988",
    abstract = "For a conversation to help and support, speakers should maintain an {\textquotedblleft}effect-effort{\textquotedblright} tradeoff. As outlined in the gist of {\textquotedblleft}Cognitive Relevance Principle{\textquotedblright}, helpful speakers should optimize the {\textquotedblleft}cognitive relevance{\textquotedblright} through maximizing the {\textquotedblleft}cognitive effects{\textquotedblright} and minimizing the {\textquotedblleft}processing effort{\textquotedblright} imposed on listeners. Although preference learning methods have given rise a boon of studies in pursuit of{\textquotedblleft}effect-optimization{\textquotedblright}, none have delved into the critical {\textquotedblleft}effort-optimiazation{\textquotedblright} to fully cultivate the awareness of {\textquotedblleft}optimal relevance{\textquotedblright} into thecognition of conversation agents. To address this gap, we integrate the {\textquotedblleft}Cognitive Relevance Principle{\textquotedblright} into emotional support agents in the environment of multi-turn conversation. The results demonstrate a significant and robust improvement against the baseline systems with respect to response quality, human-likedness and supportivenss. This study offers compelling evidence for the effectiveness of the {\textquotedblleft}Relevance Principle{\textquotedblright} in generating human-like, helpful, and harmless emotional support conversations. The source code will be available at https://github.com/CN-Eyetk/VLESA-ORL.git"
}