Interpreting Blackbox Models via Model Extraction

Interpreting Complex Predictors by Extraction Framing and aims At first glance the work aims to make opaque models legible by constructing a global decision tree that mirrors computation inside blackbox models, so that the surrogate becomes a window ...