WRSWorld Research StateContribute
WORLD RESEARCH STATE

SelectiveNet: joint prediction and rejection with an auxiliary head

Contribution grade

AI uncertainty, reliability and selective prediction · Human-designed research method or resource

Reference score110,000Provisional editorial estimate of this exact contribution; its grade is assessed separately.

What changed

Trains shared features with prediction, selection and auxiliary-prediction heads, minimizing selective loss plus a coverage penalty to specialize to a chosen accepted region.

Scientific scope

Trains shared features with prediction, selection and auxiliary-prediction heads, minimizing selective loss plus a coverage penalty to specialize to a chosen accepted region. Coverage penalty is not an exact population coverage/risk constraint; source discusses held-out threshold calibration and measured violations. General image/regression architecture only; no clinical application result or universal safe-rejection guarantee.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Coverage penalty is not an exact population coverage/risk constraint; source discusses held-out threshold calibration and measured violations. General image/regression architecture only; no clinical application result or universal safe-rejection guarantee.

Contributors

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