arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2504.20733cs.LGcs.AI

无监督代理异常检测

Unsupervised Surrogate Anomaly Detection

Department of Computer Science TU Dortmund University(计算机科学系多特蒙德大学)

Simon Klüttermann, Tim Katzke, Emmanuel Müller

AI总结:

本文提出代理异常检测方法论,将其形式化为最优代理模型公理并设计DEAN算法,在121个基准数据集上与19种现有方法相比表现出竞争性能,验证了可扩展性和可靠性。

摘要:

本文研究无监督异常检测算法,这类算法学习神经网络表示(即正常数据的规律模式),异常数据会偏离该模式。受工程领域类似概念启发,我们将方法论称为代理异常检测。我们将代理异常检测概念形式化为最优代理模型所需的一组公理,并提出新算法DEAN(深度集成异常检测),其设计旨在满足这些标准。我们在121个基准数据集上评估DEAN,结果表明其相对于19种现有方法具有竞争性能,同时验证了我们方法的可扩展性和可靠性。

英文摘要:

In this paper, we study unsupervised anomaly detection algorithms that learn a neural network representation, i.e. regular patterns of normal data, which anomalies are deviating from. Inspired by a similar concept in engineering, we refer to our methodology as surrogate anomaly detection. We formalize the concept of surrogate anomaly detection into a set of axioms required for optimal surrogate models and propose a new algorithm, named DEAN (Deep Ensemble ANomaly detection), designed to fulfill these criteria. We evaluate DEAN on 121 benchmark datasets, demonstrating its competitive performance against 19 existing methods, as well as the scalability and reliability of our method.

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。

↑