COVER: Identifiable Evaluation of Coalition Routing· 联合路由的可识别评估
When a multi-agent system changes its team, it also changes the messages and final answer it produces, so an end-to-end accuracy gap does not by itself identify a routing effect. We introduce method, an evaluation contract that fixes a public information boundary, downstream stack G, and finite legal team family before outcomes are generated. Complete coverage identifies exact finite-benchmark oracle regret conditional on that stack. For any finite collection of frozen policies, executing the union of their distinct selected teams is the minimal assumption-free support for every pairwise policy contrast, though not for absolute oracle regret. Two controlled tables with source-ID-disjoint splits test the instrument. On MuSiQue-12, a pre-specified privileged positive control improves regret
提出联合路由可识别评估方法，提高多智能体系统性能评估准确性。
- 核心方法
- 引入评估合同方法，通过固定公共信息边界、下游堆栈和有限合法团队家族，实现对团队变化前后的精确评估。
- 适合谁读
- 研究者
- 要解决的问题
- 多智能体系统改变团队时产生的端到端准确度差距无法单独识别路由效果的问题。
- 关键实验
- 在MuSiQue-12数据集上进行了实验，预设的特权正控组改善了遗憾值。
- 主要贡献
- 提供了一种在多智能体系统中评估联合路由效应的可识别方法，有助于更准确地分析性能改进。
- 意义与局限
- 该方法为多智能体系统中的路由评估提供了新的视角，但可能仅适用于特定类型的策略对比而非绝对的遗憾值评估。