Counterfactual explanations may be invalidated or require more intervention than originally expected, if adverse events take place (e.g., genetic predisposition to resist a treatment). Can we search for counterfactual explanations that are robust to such unfortunate circumstances? What is the extra cost that robust counterfactual explanations entail? Is this cost justified?
Virgolin, M. (2023). marcovirgolin /robust-counterfactuals.
|View at GitHub