Behavioral Privacy Leakage in Agentic Negotiation: Formalizing and Mitigating Inference Attacks via Randomized Policies
代理协商中的行为隐私泄露:通过随机策略形式化和减轻推理攻击
专题命中 多智能体 :agentic(title)
AI总结 研究自主协商代理行为隐私泄露问题,设计自适应随机协商策略,联合保证差分隐私、报价序列收敛和高协商效用,在合成双边协商实验中降低对手推理准确率,保持高协商成功率和效用。
Comments 3 figures. Accepted at the AI4TCI Workshop (AI for Secure and Trustworthy Critical Infrastructure Systems), co-located with ARES 2026