智能体间信任评估数据集

本数据集聚焦多智能体系统中智能体间的信任评估与声誉管理，覆盖对等服务协作、任务委托、信息推荐、跨域认证、市场交易等场景。数据内容包括智能体交互历史（行为类型、结果质量、完成状态）、直接信任评分、间接信任评分（基于推荐聚合）、综合信任评分、推荐来源及权重、信任验证标签（后续实际表现）及恶意智能体标识。适用于动态信任模型训练、恶意行为检测、推荐聚合算法评测、声誉系统优化及去中心化信任基础设施验证等场景。This dataset focuses on trust evaluation and reputation management among agents in multi-agent systems, covering scenarios including peer-to-peer (P2P) service collaboration, task delegation, information recommendation, cross-domain authentication, market transactions, etc. The dataset includes agent interaction history (behavior type, result quality, completion status), direct trust scores, indirect trust scores (based on recommendation aggregation), comprehensive trust scores, recommendation sources and their weights, trust verification labels (based on subsequent actual performance), and malicious agent identifiers. It is applicable to scenarios such as dynamic trust model training, malicious behavior detection, recommendation aggregation algorithm evaluation, reputation system optimization, and decentralized trust infrastructure validation.

搜集汇总

背景与挑战

背景概述

该数据集聚焦多智能体系统中智能体间的信任评估与声誉管理，覆盖服务协作、任务委托等多种场景，包含交互历史、信任评分及验证标签等结构化数据。它适用于动态信任模型训练、恶意行为检测、推荐算法评测及声誉系统优化等应用。