1 paper · 1 filter
Khawla Elhadri, Tomasz Michalski, Adam Wróbel +3
The growing interest in eXplainable Artificial Intelligence (XAI) has stimulated research on models with built-in interpretability, among which part-prototype models are particular…