UnderstandingGlobalFeatureContributionsWith AdditiveImportanceMeasures
–Neural Information Processing Systems
Most recent research hasaddressed thisby focusing onlocal interpretability, which explains a model's individual predictions (e.g., the role of each feature in a patient's diagnosis) [25, 30, 34, 38]. Twospecial cases areS = andS = D, which respectively correspond to the mean prediction f (x ) = E[f(X)] and the full model predictionfD(x) = f(x).
Neural Information Processing Systems
Feb-10-2026, 07:33:32 GMT