Improving Efficiency and Accuracy of Causal Discovery Using a Hierarchical Wrapper

TL;DR

This paper presents a recursive wrapper strategy for causal discovery algorithms that reduces the number of statistical tests needed, improves accuracy, and shortens runtime by clustering variables and learning local causal graphs.

Contribution

It introduces a novel hierarchical wrapper that enhances existing causal discovery methods by clustering variables and efficiently combining local graphs while maintaining soundness and completeness.

Findings

Requires fewer statistical tests than baseline algorithms.

Learns more accurate causal graphs in synthetic and real-world data.

Reduces runtime significantly compared to traditional methods.

Abstract

Causal discovery from observational data is an important tool in many branches of science. Under certain assumptions it allows scientists to explain phenomena, predict, and make decisions. In the large sample limit, sound and complete causal discovery algorithms have been previously introduced, where a directed acyclic graph (DAG), or its equivalence class, representing causal relations is searched. However, in real-world cases, only finite training data is available, which limits the power of statistical tests used by these algorithms, leading to errors in the inferred causal model. This is commonly addressed by devising a strategy for using as few as possible statistical tests. In this paper, we introduce such a strategy in the form of a recursive wrapper for existing constraint-based causal discovery algorithms, which preserves soundness and completeness. It recursively clusters the observed variables using the normalized min-cut criterion from the outset, and uses a baseline causal discovery algorithm during backtracking for learning local sub-graphs. It then combines them and ensures completeness. By an ablation study, using synthetic data, and by common real-world benchmarks, we demonstrate that our approach requires significantly fewer statistical tests, learns more accurate graphs, and requires shorter run-times than the baseline algorithm.