CausalKANs: interpretable treatment effect estimation with Kolmogorov-Arnold networks

2025-09-26

AI Summary

To address the limited interpretability of deep neural networks in estimating heterogeneous treatment effects (CATE) for high-stakes domains, this paper introduces the first causal inference framework incorporating Kolmogorov–Arnold Networks (KANs). Our method jointly optimizes KAN reconstruction, structured pruning, and symbolic regression to transform black-box CATE models into auditable closed-form analytical expressions, accompanied by interpretable visualizations. Crucially, it preserves neural-network-level predictive accuracy while substantially enhancing model transparency: on multiple benchmark datasets, its CATE estimation error matches that of state-of-the-art causal neural networks—even lightweight KAN variants remain competitive. The core contribution is the first unification of high-accuracy CATE estimation with rigorous mathematical interpretability, establishing a verifiable and deployable paradigm for trustworthy decision-making in domains such as medicine and economics.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Deep neural networks achieve state-of-the-art performance in estimating heterogeneous treatment effects, but their opacity limits trust and adoption in sensitive domains such as medicine, economics, and public policy. Building on well-established and high-performing causal neural architectures, we propose causalKANs, a framework that transforms neural estimators of conditional average treatment effects (CATEs) into Kolmogorov--Arnold Networks (KANs). By incorporating pruning and symbolic simplification, causalKANs yields interpretable closed-form formulas while preserving predictive accuracy. Experiments on benchmark datasets demonstrate that causalKANs perform on par with neural baselines in CATE error metrics, and that even simple KAN variants achieve competitive performance, offering a favorable accuracy--interpretability trade-off. By combining reliability with analytic accessibility, causalKANs provide auditable estimators supported by closed-form expressions and interpretable plots, enabling trustworthy individualized decision-making in high-stakes settings. We release the code for reproducibility at https://github.com/aalmodovares/causalkans .

Problem

• Estimating interpretable treatment effects using Kolmogorov-Arnold networks
• Transforming neural CATE estimators into interpretable closed-form formulas
• Balancing predictive accuracy with interpretability for trustworthy decision-making

Innovation

• Uses Kolmogorov-Arnold Networks for treatment effects
• Applies pruning and symbolic simplification techniques
• Generates interpretable closed-form formulas