研究人员推出SolarChain-Eval，一个旨在评估在去中心化能源市场中运行的AI代理可信度的新基准。该基准纳入了物理约束，用于评估代理在市场效用之外的指标，包括物理安全性、滑点和可审计性。实验表明效用和安全性之间存在权衡，强化学习代理提高了效用但可能表现出不安全行为。基于LLM的规划/审计层可以提高可审计性并减轻一些风险，但不能完全弥补奖励函数定义不当的问题。 AI
影响 该基准可能导致在能源市场等关键基础设施中出现更可靠、更安全的AI应用。
排序理由 该集群包含一篇介绍AI代理新基准的研究论文。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
arXiv:2607.08681v1 Announce Type: new Abstract: As agentic AI systems are increasingly applied to cyber-physical environments, their evaluation requires assessment of both task performance and trustworthiness. In decentralized energy markets, autonomous agents may improve market …
As agentic AI systems are increasingly applied to cyber-physical environments, their evaluation requires assessment of both task performance and trustworthiness. In decentralized energy markets, autonomous agents may improve market utility, but may also exploit invalid physical d…