多智能体系统在何处失败？集体机制的证据基础诊断

中文摘要 AI 辅助

当一个多智能体系统给出正确答案时，人们很容易得出结论：其智能体按预期共享、核查并使用了信息。然而，系统可能破坏其某一集体机制（即管理智能体如何路由、接纳、存储及处理共享信息的规则），却仍能返回正确答案；而错误答案很少能揭示哪个机制失效。我们提出一个问题：何种执行证据足以断定某一特定机制被违反？我们的答案是诊断契约（diagnostic contract），它将什么构成违规与哪些执行记录能确立违规区分开来，并得出结论：违规得到支持、被排除或未知；删除记录可能使该结论变为未知，但绝不会使其反转。我们针对四种机制测试了契约，通过从机制作用的步骤开始重放执行，一次保持不变，一次机制被破坏，一次机制被恢复。被破坏的机制常常仍留下正确答案。一个LLM诊断器从内部记录中检测到的违规远多于从公开输出中检测到的，然而在相同记录下，通用提示词常声称记录并不支持的确定性，而陈述契约的提示词在很大程度上避免了这一点。这些契约还以狭义形式适用于独立开发系统中的机制，但一种在我们基准上近乎完美的诊断行为在独立开发的工作流上却性能下降。因此，正确结果不能替代关于集体机制如何运作的记录，且在一个基准上的共识并不能证明诊断器能迁移到另一个系统。

英文摘要

When a multi-agent system answers correctly, it is tempting to conclude that its agents shared, checked, and used information as intended. Yet a system can break one of its collective mechanisms, the rules that govern how agents route, admit, store, and act on shared information, and still return the right answer, while a wrong answer rarely reveals which mechanism failed. We ask what evidence from an execution is sufficient to conclude that a particular mechanism was violated. Our answer is a diagnostic contract, which separates what counts as a violation from which execution records can establish one, and concludes that a violation is supported, ruled out, or unknown; removing records can make this conclusion unknown but never reverse it. We test contracts for four mechanisms by replaying executions from the step where a mechanism acts, once unchanged, once with the mechanism broken, and once with it restored. Broken mechanisms often left the answer correct. An LLM diagnoser detected many more violations from internal records than from public outputs, yet with identical records a generic prompt often claimed certainty the records did not support, which prompts stating the contracts largely avoided. The contracts also applied, in narrow form, to mechanisms in independently developed systems, but a diagnostic behavior that was nearly perfect on our benchmark degraded on an independently developed workflow. A correct outcome is therefore no substitute for records of how collective mechanisms operated, and agreement on one benchmark does not show that a diagnoser transfers to another system.

发表机构

• New York University（纽约大学）

机构由 AI 辅助整理，请以论文原文为准。