Outlier Exposure: anomaly detection trained with auxiliary outlier data

What changed

Trains a representation or detector with a complementary outlier loss on a diverse auxiliary dataset disjoint from the test anomalies, improving selected image and text anomaly detection.

Scientific scope

Trains a representation or detector with a complementary outlier loss on a diverse auxiliary dataset disjoint from the test anomalies, improving selected image and text anomaly detection. Generalization to unseen anomalies is empirically demonstrated, not universal. Choice and overlap of auxiliary data matter; original methodology differs from methods tuned to each tested OOD source and results are not automatically comparable.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Generalization to unseen anomalies is empirically demonstrated, not universal. Choice and overlap of auxiliary data matter; original methodology differs from methods tuned to each tested OOD source and results are not automatically comparable.

Contributors

• Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.

• Deep Anomaly Detection with Outlier Exposure

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction