Library · Artifacts · Fair for Whom?
Artifact

Fair for Whom?

Fairness reframed as validity asked one subgroup at a time. An aggregate accuracy number can look fine while the model quietly degrades for smaller groups — differential prediction hiding under the average.