매일 아침, 어제의 AI를 한 통으로 정리해 보내드립니다메일로 받아보기

METAL LAB

Certify or Refuse: A Cross-Model Map for Selective Risk Control with Coverage Floors under Covariate Shift

arXiv:2608.108932026-08-12

arXiv:2608.10893v1 Announce Type: new Abstract: Certified selective predictors attain whatever coverage they attain; operators impose an automation floor: answer at least a $\beta$-fraction of shifted target traffic with at most an $\alpha$-fraction of answers wrong. Under bounded-ratio covariate shift we prove the Floor Certification Map: once that floor must be certified alongside the selection-conditioned risk $\alpha$, certification acquires a feasibility frontier and a two-resource complexi

저자 · Jiamiao Liu, Dewen Qiao, Yu Zhang, Xuetao Chen

arXiv에서 원문 보기