Software
Software
COunterfactual explanations with Limited Actions (COLA)
COunterfactual explanations with Limited Actions (COLA)
Explainable Artificial Intelligence (XAI) is essential for making artificial intelligence systems transparent and trustworthy. Within this area, feature attributions (FA) methods, such as Shapley values determine how much each input feature contributes to a machine learning (ML) model's output. Another technique counterfactual explanations (CE) show how small changes in input features can lead to different outcomes. COLA adapts to various CE methods and ML models. Specifically, given a group of observed instance and a desired target model outcome, COLA refines the observations to yield a counterfactual explanations within a pre-defined maximum number of feature modifications.