Concept拒识学习 / 分诊框架
Selective Classification
A machine learning framework that allows a model to abstain from prediction when confidence is insufficient and to refer the sample to human review or a secondary system. Its core evaluation criterion is the coverage-accuracy trade-off.
Timeline (last 90 days)
Oct 20
In the selective classification framework, when model confidence is below the threshold τ=0.934, the sample is rejected and referred to an expert for review.
Unverified50%