1 paper
Xinmiao Lin, Matthew Wright
Prototype-based neural networks provide inherent interpretability through case-based reasoning, yet suffer from critical limitations: prototypes converge to redundant features, fai…