1 paper · 1 filter
Stefano Fioravanti, Francesco Giannini, Paolo Frazzetto +2
The most common methods in explainable artificial intelligence are post-hoc techniques which identify the most relevant features used by pretrained opaque models. Some of the most…