Practical do-Shapley Explanations with Estimand-Agnostic Causal Inference

2025-09-24

AI Summary

SHAP ignores causal structure, while do-SHAP relies on a pre-specified estimand, limiting its practical applicability. Method: We propose an estimand-agnostic causal SHAP framework that dispenses with assumptions about predefined causal parameters and unifies the modeling of arbitrary identifiable interventional queries. Our approach integrates interventional SHAP, causal graph models, and approximation-based optimization algorithms to jointly estimate diverse causal effects—including total, direct, and indirect effects—within a single model, while also enabling interpretation of latent-variable-driven data-generating processes. Contribution/Results: Evaluated on two real-world datasets, our method achieves up to 3.2× speedup in computation while maintaining high-fidelity causal attribution. It advances do-SHAP toward a general-purpose, practical, and scalable framework for causal explainable AI.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Humans and AI: Explainable AI (XAI) for Human Understanding

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization

Abstract

Among explainability techniques, SHAP stands out as one of the most popular, but often overlooks the causal structure of the problem. In response, do-SHAP employs interventional queries, but its reliance on estimands hinders its practical application. To address this problem, we propose the use of estimand-agnostic approaches, which allow for the estimation of any identifiable query from a single model, making do-SHAP feasible on complex graphs. We also develop a novel algorithm to significantly accelerate its computation at a negligible cost, as well as a method to explain inaccessible Data Generating Processes. We demonstrate the estimation and computational performance of our approach, and validate it on two real-world datasets, highlighting its potential in obtaining reliable explanations.

Problem

• SHAP explanations often ignore causal problem structure
• do-SHAP's reliance on estimands limits practical application
• Need to make causal explanations feasible on complex graphs

Innovation

• Estimand-agnostic causal inference for do-SHAP
• Novel algorithm accelerates computation significantly
• Method explains inaccessible Data Generating Processes