Searching for actual causes: Approximate algorithms with adjustable precision

TL;DR

This paper introduces polynomial-time algorithms with adjustable precision to identify actual causes in complex systems, addressing limitations of existing methods and enhancing explainability for non-expert users.

Contribution

It proposes novel approximate algorithms with tunable precision for identifying actual causes, applicable to diverse and complex system types.

Findings

Algorithms identify causes in non-boolean, black-box, and stochastic systems.

Adjustable algorithms balance precision and computational effort.

Experiments demonstrate effectiveness over existing approaches.

Abstract

Causality has gained popularity in recent years. It has helped improve the performance, reliability, and interpretability of machine learning models. However, recent literature on explainable artificial intelligence (XAI) has faced criticism. The classical XAI and causality literature focuses on understanding which factors contribute to which consequences. While such knowledge is valuable for researchers and engineers, it is not what non-expert users expect as explanations. Instead, these users often await facts that cause the target consequences, i.e., actual causes. Formalizing this notion is still an open problem. Additionally, identifying actual causes is reportedly an NP-complete problem, and there are too few practical solutions to approximate formal definitions. We propose a set of algorithms to identify actual causes with a polynomial complexity and an adjustable level of precision and exhaustiveness. Our experiments indicate that the algorithms (1) identify causes for different categories of systems that are not handled by existing approaches (i.e., non-boolean, black-box, and stochastic systems), (2) can be adjusted to gain more precision and exhaustiveness with more computation time.