Improving Local Explainability By Learning Causal Graphs From Data

Causal Shapley values take into account causal relations among dependent features to adjust the contributions of each feature to a prediction. A limitation of this approach is that it can only leverage known causal relations. In this work we combine the computation of causal Shapley values with causal discovery, i.e., learning causal graphs from data. In particular, we compute causal explanations across the Markov Equivalence Class (MEC), a set of candidate causal graphs learned from observational data, providing a list of causal Shapley values that explain the prediction. We propose two methods for estimating this list efficiently, drawing on the equivalences of the interventional distributions for a subset of the causal graphs. We evaluate our methods on synthetic and real-world data, showing that they provide explanations that are more consistent with the true causal effects compared to traditional Shapley value approaches that disregard causal relations. Our results show that even when the Markov Equivalence Class is learned incorrectly, in most settings the explanations of our framework are on average closer to true causal Shapley values than marginal and conditional Shapley values.

Authors

• Daan Roos
• Sebastian Gerwinn
• Jan-Willem van de Meent
• Sara Magliacane

Keywords

No keywords are indexed for this paper.