Natural Counterfactuals With Necessary Backtracking
This addresses the challenge of generating realistic counterfactual explanations for decision-making, though it appears incremental by refining an existing approach.
The paper tackles the problem that Pearl's counterfactual reasoning often requires unrealistic deviations from observed scenarios, and proposes a framework of natural counterfactuals with backtracking to generate more feasible scenarios, showing effectiveness in empirical experiments.
Counterfactual reasoning is pivotal in human cognition and especially important for providing explanations and making decisions. While Judea Pearl's influential approach is theoretically elegant, its generation of a counterfactual scenario often requires too much deviation from the observed scenarios to be feasible, as we show using simple examples. To mitigate this difficulty, we propose a framework of \emph{natural counterfactuals} and a method for generating counterfactuals that are more feasible with respect to the actual data distribution. Our methodology incorporates a certain amount of backtracking when needed, allowing changes in causally preceding variables to minimize deviations from realistic scenarios. Specifically, we introduce a novel optimization framework that permits but also controls the extent of backtracking with a naturalness criterion. Empirical experiments demonstrate the effectiveness of our method. The code is available at https://github.com/GuangyuanHao/natural_counterfactuals.