On The Stability of Interpretable Models

Riccardo Guidotti, Salvatore Ruggieri

Interpretable classification models are built with the purpose of providing a comprehensible description of the decision logic to an external oversight agent. When considered in is…