We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Anomaly Detection by Recombining Gated Unsupervised Experts

Abstract: Anomaly detection has been considered under several extents of prior knowledge. Unsupervised methods do not require any labelled data, whereas semi-supervised methods leverage some known anomalies. Inspired by mixture-of-experts models and the analysis of the hidden activations of neural networks, we introduce a novel data-driven anomaly detection method called ARGUE. Our method is not only applicable to unsupervised and semi-supervised environments, but also profits from prior knowledge of self-supervised settings. We designed ARGUE as a combination of dedicated expert networks, which specialise on parts of the input data. For its final decision, ARGUE fuses the distributed knowledge across the expert systems using a gated mixture-of-experts architecture. Our evaluation motivates that prior knowledge about the normal data distribution may be as valuable as known anomalies.
Comments: Accepted at IJCNN 2022
Subjects: Machine Learning (cs.LG); Machine Learning (stat.ML)
Cite as: arXiv:2008.13763 [cs.LG]
  (or arXiv:2008.13763v5 [cs.LG] for this version)

Submission history

From: Jan-Philipp Schulze [view email]
[v1] Mon, 31 Aug 2020 17:35:57 GMT (754kb,D)
[v2] Mon, 12 Oct 2020 15:58:25 GMT (387kb,D)
[v3] Sun, 4 Apr 2021 16:58:17 GMT (156kb)
[v4] Sun, 6 Mar 2022 13:09:57 GMT (143kb,D)
[v5] Mon, 23 May 2022 17:02:34 GMT (148kb,D)

Link back to: arXiv, form interface, contact.