While reading about model explainability and model accountability, the term surrogate model keeps appearing. I had an idea about what it is but it does not seem to make sense anymore:
What is a surrogate models?
Why are surrogate models good for explainability?