Partially Interpretable Decision Trees

Ludovic Schwartz, Federico Berton, Tim van Erven

Partially interpretable models extend simple interpretable models with the option to defer complicated inputs to an accurate but uninterpretable black-box model. By deferring more…