Multi-agent collaborative crowd intelligence persuasive dialogue inspired by cognitive psychology

Persuasion is a pivotal skill in human interaction, and endowing human-machine dialogue systems with robust persuasive capabilities can effectively control the direction and subjective logic of interactions, thus fostering further intelligence in human-machine interactions.Current research in persuasive human-machine dialogue lacks a model of the persuasive mechanism itself and focuses only on simple one-on-one persuasive scenarios, which do not adapt well to complex and more universally applicable crowd persuasion contexts. In light of this, this paper introduces the concept of “crowd intelligence persuasion dialogue," employing a variety of persuasive agents with distinct roles. These agents collaboratively employ diversified crowd persuasion strategies from different topical perspectives to efficiently accomplish persuasion tasks.Initially, this study begins with foundational theories in cognitive psychology to summarize and deduce dialogue strategies of crowd intelligence persuasion that can effectively alter the persuader's cognitive perceptions. Following this, a deep analysis of the persuasive process in complex crowd interaction scenarios is conducted, establishing a general system architecture for crowd intelligence persuasion dialogue tailored to intricate scenarios. This facilitates a comprehensive understanding and modeling of crowd intelligence persuasion dialogue scenarios.Subsequently, this paper discusses the research challenges and key technologies in the field of crowd intelligence persuasion dialogue and concludes with an overview of our research practices in this area.