CASE: Contrastive Activation for Class-Sensitive Explanations
This work proposes a diagnostic test for class sensitivity: a method’s ability to distinguish between competing class labels on the same input, and introduces CASE, a contrastive explanation method that isolates features uniquely discriminative for the predicted class.
Dane Williamson, Yang-Feng Ji, Matthew B. Dwyer
· Machine-mediated learning · 1 citation