#intrinsic-interpretability

Question 1: Single choice

A hospital replaces a deep network, which it explained with saliency maps, by a sparse logistic regression whose few coefficients doctors can read directly. How is this change best described?

From post-hoc explanation to an interpretable-by-design model1
From a local post-hoc method to a global post-hoc method2
From a model-agnostic to a model-specific explanation method3
From a faithful explanation to a merely plausible explanation4

#intrinsic-interpretability

Timed test