Unsupervised Surrogate Anomaly Detection

Simon Klüttermann, Tim Katzke, Emmanuel Müller · 2025

Read Unsupervised Surrogate Anomaly Detection by Simon Klüttermann, Tim Katzke, Emmanuel Müller (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-20733.txt.

Description

Simon Kluttermann, Tim Katzke, and Emmanuel Muller study unsupervised anomaly detection in a 2025 paper. Such methods learn the normal pattern in data, so that odd points stand out as deviations. The authors borrow an engineering idea and call the learned normal model a surrogate.

They first state a set of axioms that an ideal surrogate should meet. They then propose DEAN, a deep ensemble method designed to meet those rules. Tests covered more than one hundred benchmark datasets against many rival methods. DEAN performed competitively and scaled well. The paper offers both a clearer theory of the task and a practical new algorithm.

Catalog record

CREATORSimon Klüttermann, Tim Katzke, Emmanuel Müller
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE79 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-20733

Browse the full library · About the Tower of Babel