给AI多智能体系统加了个质量信号标签，能自动区分该修复还是停止。
Trustworthy Multi-Agent Systems: Mitigating Semantic Drift with the Argent Signaling Protocol
- 给每条AI回答加四个质量信号，判断是否可信
- 在3个模型上提升通过率，最高达44.4%
- 能100%拦截无依据的错误信息传播
当多智能体大模型系统给出错误答案时，并非所有失败都一样：有些是信息不全但基于正确材料，有些则是完全无依据，应立即终止。现有重试策略对两者一视同仁（重试并碰运气），导致人工无法判断是否该重试或直接停止。本文提出Argent信号协议（ASP），一种紧凑的机器可读头部，为每条AI生成的回答附带结构化质量信号：确定性（@C）、依据性（@G）、随机性（@S）和假设指数，用于分类每个断言的证据基础。这些信号使控制器能区分可修复失败与需遏制失败，并分别处理。我们在两种模式下评估了ASP。在独立模式下，使用27题文档问答基准测试（基于Array BioPharma/Ono许可协议），对比基线提示与ASP控制动作在三个本地GGUF模型上的表现。在Qwen（0.8B）上，通过率从11.1%提升至33.3%，平均术语覆盖率从36.7%升至65.4%；在Dobby（8B）上，实现4次‘失败转成功’，通过率从33.3%增至44.4%；在SmolLM3（3B）上，按题交替采用修复与遏制策略。总体通过数从12/81提升至21/81。在多智能体模式中，一个ASP旁路组件位于检索代理与下游决策代理之间，成功阻断100%无依据的上游输出（24/27被拦截，0个错误传播至下游）。
原文摘要 · Abstract (English)
When multi-agent LLM systems produce bad answers, not all failures are equal: some answers are grounded in the right material but incomplete, while others are simply ungrounded and should be stopped. Current retry strategies treat both cases identically (try again and hope for the best), leaving human supervisors unable to tell whether a retry was warranted or whether the system should have halted instead. We introduce the Argent Signaling Protocol (ASP), a compact machine-readable header that accompanies every AI-generated response with structured quality signals: certainty (@C), grounding (@G), stochasticity (@S), and an assumption index that classifies the evidentiary basis of each claim. These signals enable a controller to distinguish repairable failures from containment failures and route each case differently. We evaluate ASP in two modes. In standalone mode, a 27-question document-grounded QA benchmark over the Array BioPharma/Ono license agreement compares baseline prompts against ASP-instrumented controller actions across three local GGUF models. On Qwen~(0.8B), ASP improves pass rate from 11.1% to 33.3% and mean term coverage from 36.7% to 65.4%; on Dobby~(8B), ASP produces 4 fail-to-pass recoveries, raising pass rate from 33.3% to 44.4%; on SmolLM3~(3B), ASP alternates between repair and containment per question. Aggregate improvement is meaningful (12/81 to 21/81 passes). In multi-agent mode, an ASP sidecar sits between a retrieval agent and a downstream decision agent; the sidecar blocks 100% of ungrounded upstream outputs from reaching the downstream agent (24/27 blocked, 0 ungrounded propagations).
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。