arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2511.04032cs.AI
收藏

检测多智能体人工智能轨迹中的静默故障

Detecting Silent Failures in Multi-Agentic AI Trajectories

IBM Research（IBM研究院）；IIIT Bangalore（班加罗尔IIIT）

Divya Pathak, Harshit Kumar, Anuska Roy, Felix George, Mudit Verma, Pratibha Moogi

阅读 PDF
arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文首次系统研究多智能体AI轨迹中的静默故障检测，提出数据集构建流程并构建两个基准数据集，实验表明XGBoost和SVDD分别达到98%和96%的准确率。

摘要：

由大型语言模型（LLMs）驱动的多智能体人工智能系统本质上具有非确定性，容易出现静默故障，如漂移、循环和输出中缺失细节，这些故障难以检测。我们引入了智能体轨迹中的异常检测任务，以识别这些故障，并提出了一种数据集构建流程，该流程能够捕获用户行为、智能体非确定性以及LLM变化。利用该流程，我们构建并标注了两个基准数据集，分别包含来自多智能体人工智能系统的4,275条和894条轨迹。在这些数据集上对异常检测方法进行基准测试，我们表明监督方法（XGBoost）和半监督方法（SVDD）表现相当，分别达到最高98%和96%的准确率。这项工作提供了对多智能体人工智能系统中异常检测的首次系统性研究，提供了数据集、基准和见解，以指导未来研究。

英文摘要：

Multi-Agentic AI systems, powered by large language models (LLMs), are inherently non-deterministic and prone to silent failures such as drift, cycles, and missing details in outputs, which are difficult to detect. We introduce the task of anomaly detection in agentic trajectories to identify these failures and present a dataset curation pipeline that captures user behavior, agent non-determinism, and LLM variation. Using this pipeline, we curate and label two benchmark datasets comprising \textbf{4,275 and 894} trajectories from Multi-Agentic AI systems. Benchmarking anomaly detection methods on these datasets, we show that supervised (XGBoost) and semi-supervised (SVDD) approaches perform comparably, achieving accuracies up to 98% and 96%, respectively. This work provides the first systematic study of anomaly detection in Multi-Agentic AI systems, offering datasets, benchmarks, and insights to guide future research.

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。

↑