{"type":"ScholarlyArticle","title":"PersuHMM: Iterative Learning the Hierarchical Meta-Strategy Memory for Persuasive Dialogue","authors":[{"name":"Yanyue Zhang","slug":"yanyue-zhang","given":"Yanyue","family":"Zhang","affiliation":""},{"name":"Zihao Wang","slug":"zihao-wang","given":"Zihao","family":"Wang","affiliation":""},{"name":"Xin Zhang","slug":"xin-zhang","given":"Xin","family":"Zhang","affiliation":""},{"name":"Deyu Zhou","slug":"deyu-zhou","given":"Deyu","family":"Zhou","affiliation":""}],"abstract":"Persuasive dialogue aims to alter people's attitudes or behaviors through conversation. \r\nWhile LLMs can generate emotional responses, they tend to use a uniform approach across varied scenarios, limiting their persuasive effectiveness.\r\nDiverse demands need varied persuasion strategies, which challenge both LLMs and human efforts.\r\nTo address this, this paper incorporates the idea of meta-learning and proposes PersuHMM, an iterative learning approach for constructing the Hierarchical Meta-strategy Memory to generate Persuasive dialogue.\r\nThrough the interaction between Meta‑layer and Task‑layer, the approach continuously accumulates automated meta‑strategies from generated outcomes and data labels, forming a persuasive meta‑strategy memory that balances efficiency and generalizability.\r\nTo facilitate model retrieval and use, the meta-strategy memory is hierarchically organized via clustering for a clearer and more logical storage structure.\r\nExperimental results show that the hierarchical meta-strategy memory enhances both the persuasiveness and generalizability of the persuasion model, and even enables cross-model performance improvement.","doi":"10.24963/ijcai.2026/677","url":"https://ijcai.org/proceedings/2026/677","pdf":"https://ijcai.org/proceedings/2026/0677.pdf","markdown":"https://ijcai.org/proceedings/2026/677.md","bibtex":"https://ijcai.org/proceedings/2026/bibtex/677","volume":{"slug":"2026","title":"Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2026","url":"https://ijcai.org/proceedings/2026/","year":2026,"location":"Bremen, Germany","dates":"15-21 August 2026","editors":"Diego Calvanese","isbn":"978-1-956792-09-6","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2026"},"section":"Natural Language Processing","track":"Main Track","number":677,"pages":{"first":6082,"last":6090},"keywords":["Natural Language Processing: Language generation","Natural Language Processing: Dialogue and interactive systems","Natural Language Processing: Applications"],"videos":[],"references":[],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}