arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2606.02494cs.SEcs.AI
收藏

在代理系统可靠之前对其进行监控

Monitoring Agentic Systems Before They're Reliable

Reins AI USA(Reins AI美国公司);Veraitech USA(Veraitech美国公司)

Marisa Ferrara Boston, Glen Hanson, Effi Georgala, JD Hudgens, Heather Frase

AI总结:

针对生产环境中代理系统因结构缺陷主导故障的问题,提出一种基于方差信号的三维度三范围监控与分类方法,并通过合成测试验证其有效性。

摘要:

进入生产环境的代理系统通常以部分集成的组件形式运行,其中结构缺陷(而非任务级错误)主导故障场景。在此成熟度下,任务级错误检测可能不可行:结构故障模式掩盖了任务级监控器旨在检测的信号。我们提出一种监控与分类方法,将代理系统评估分解为三个维度(质量、适用性、效率)和三个监控范围(运行内、跨运行、结构),使用方差作为表征信号。发现结果通过基于FMEA的严重性分类进行路由,将人类注意力集中在需要调查的子集上。我们在一个包含220次运行、120个文档包且受控错误注入的合成测试平台上进行评估。三个结果显现:监控范围决定故障类型——运行内监控器发现确定性阶段缺陷(CV=0.02),跨运行监控器发现随机集成后果(CV=1.25,24%为L2级),结构监控器以完全一致性识别集成缺口(CV=0.00)。注入的任务级错误与干净基线无法区分,证实结构缺陷掩盖了任务级信号。确定性分类将97%的发现路由至自动跟踪,仅留下2%反映可变行为的发现供人工调查。基于第一阶段证据,我们提出一个成熟度阶段模型,其中监控随着集成缺陷的解决从结构表征过渡到错误检测再到可靠性跟踪。该分类法、基于CV的范围表征和严重性模型在架构上可迁移至受监管行业中基于文档的多阶段代理工作流;具体校准是领域特定的。尽早部署监控:它发现的第一个问题就是最需要修复的问题。

英文摘要:

Agentic systems entering production typically operate as partially integrated assemblies where structural defects, not task-level errors, dominate the failure landscape. At this maturity level, task-level error detection may be infeasible: structural failure modes mask the signal that task-level monitors are designed to detect.We present a monitoring and triage methodology that decomposes agentic system evaluation into three dimensions (quality, suitability, efficiency) at three monitoring scopes (within-run, cross-run, structural), using variance as a characterization signal. Findings are routed through severity classification adapted from FMEA, concentrating human attention on the subset that warrants investigation. We evaluate on a synthetic testbed of 220 runs across 120 document bundles with controlled error injection.Three results emerge. Monitor scope determines failure type: within-run monitors surface deterministic stage defects (CV = 0.02), cross-run monitors surface stochastic integration consequences (CV = 1.25, 24% at L2), and a structural monitor identifies an integration gap with perfect consistency (CV = 0.00). Injected task-level errors are indistinguishable from clean baselines, confirming structural defects mask task-level signal. Deterministic triage routes 97% of findings to automated tracking, leaving the 2% reflecting variable behavior for human investigation.We propose, on Stage 1 evidence, a maturity-staging model in which monitoring transitions from structural characterization to error detection to reliability tracking as integration defects resolve. The taxonomy, CV-based scope characterization, and severity model transfer architecturally to document-driven, multi-stage agentic workflows in regulated industries; specific calibrations are domain-specific. Deploy monitoring early: the first thing it finds is the most important thing to fix.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑