Maximum-softmax baseline for prediction errors and out-of-distribution detection

What changed

Uses maximum softmax probability to rank prediction errors and anomalous inputs, establishing cross-domain detection baselines with explicit ROC and precision-recall evaluation.

Scientific scope

Uses maximum softmax probability to rank prediction errors and anomalous inputs, establishing cross-domain detection baselines with explicit ROC and precision-recall evaluation. A simple empirical baseline, not a calibrated probability of correctness or universal OOD detector. Precision-recall depends on class prevalence; the source explicitly documents high-confidence errors and anomalies. Star marks prior affiliation, not equal contribution.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. A simple empirical baseline, not a calibrated probability of correctness or universal OOD detector. Precision-recall depends on class prevalence; the source explicitly documents high-confidence errors and anomalies. Star marks prior affiliation, not equal contribution.

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.

• A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networks

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