dcFCI: Robust Causal Discovery Under Latent Confounding, Unfaithfulness, and Mixed Data

TL;DR

dcFCI is a novel causal discovery method that robustly infers causal structures from observational data, effectively handling latent confounders, unfaithfulness, and mixed data types, outperforming existing approaches especially with small or heterogeneous datasets.

Contribution

We introduce a nonparametric score for PAG compatibility and develop dcFCI, the first hybrid algorithm addressing latent confounding, empirical unfaithfulness, and mixed data types simultaneously.

Findings

dcFCI outperforms state-of-the-art methods on synthetic and real data.

dcFCI effectively recovers true PAGs in small, heterogeneous datasets.

The proposed score accurately characterizes structural uncertainty.

Abstract

Causal discovery is central to inferring causal relationships from observational data. In the presence of latent confounding, algorithms such as Fast Causal Inference (FCI) learn a Partial Ancestral Graph (PAG) representing the true model's Markov Equivalence Class. However, their correctness critically depends on empirical faithfulness, the assumption that observed (in)dependencies perfectly reflect those of the underlying causal model, which often fails in practice due to limited sample sizes. To address this, we introduce the first nonparametric score to assess a PAG's compatibility with observed data, even with mixed variable types. This score is both necessary and sufficient to characterize structural uncertainty and distinguish between distinct PAGs. We then propose data-compatible FCI (dcFCI), the first hybrid causal discovery algorithm to jointly address latent confounding, empirical unfaithfulness, and mixed data types. dcFCI integrates our score into an (Anytime)FCI-guided search that systematically explores, ranks, and validates candidate PAGs. Experiments on synthetic and real-world scenarios demonstrate that dcFCI significantly outperforms state-of-the-art methods, often recovering the true PAG even in small and heterogeneous datasets. Examining top-ranked PAGs further provides valuable insights into structural uncertainty, supporting more robust and informed causal reasoning and decision-making.