REX: Causal discovery based on machine learning and explainability techniques
Jesus Renero, Idoia Ochoa, Roberto Maestre
TL;DR
ReX addresses the challenge of causal discovery by integrating machine-learning predictors with Shapley-value explainability to guide the reconstruction of causal graphs under the Causal Markov Condition and faithfulness. The method trains two regressor families (DFN and GBT) to predict each variable from the rest, derives SHAP-based parent sets, bootstraps to stabilize the undirected graph, orients edges via Additive Noise Models and HSIC, and resolves cycles using a SHAP-discrepancy criterion. Across five synthetic data families and real-world datasets (Sachs and finance), ReX achieves high precision and F1 while maintaining low SHD and SID, often outperforming standard causal-discovery baselines. The approach provides interpretable insights by tying feature contributions to potential causal relations and demonstrates robustness to non-linearities and multicollinearity, albeit with substantial computational cost due to SHAP. The work suggests promising directions for scalable SHAP-based causal discovery, including alternative regressors, distributional shift handling, and broader domain applications, with code and data openly available.
Abstract
Explainable Artificial Intelligence (XAI) techniques hold significant potential for enhancing the causal discovery process, which is crucial for understanding complex systems in areas like healthcare, economics, and artificial intelligence. However, no causal discovery methods currently incorporate explainability into their models to derive the causal graphs. Thus, in this paper we explore this innovative approach, as it offers substantial potential and represents a promising new direction worth investigating. Specifically, we introduce ReX, a causal discovery method that leverages machine learning (ML) models coupled with explainability techniques, specifically Shapley values, to identify and interpret significant causal relationships among variables. Comparative evaluations on synthetic datasets comprising continuous tabular data reveal that ReX outperforms state-of-the-art causal discovery methods across diverse data generation processes, including non-linear and additive noise models. Moreover, ReX was tested on the Sachs single-cell protein-signaling dataset, achieving a precision of 0.952 and recovering key causal relationships with no incorrect edges. Taking together, these results showcase ReX's effectiveness in accurately recovering true causal structures while minimizing false positive predictions, its robustness across diverse datasets, and its applicability to real-world problems. By combining ML and explainability techniques with causal discovery, ReX bridges the gap between predictive modeling and causal inference, offering an effective tool for understanding complex causal structures.