PACER: Acyclic Causal Discovery from Large-Scale Interventional Data

TL;DR

PACER is a scalable, likelihood-based framework for causal discovery from large-scale interventional data that guarantees acyclicity by construction and outperforms existing methods in speed and accuracy.

Contribution

Introduces PACER, a novel acyclicity-guaranteed causal discovery method that efficiently handles high-dimensional interventional data without surrogate penalties.

Findings

PACER matches or exceeds state-of-the-art accuracy on benchmark datasets.

Achieves up to two orders of magnitude speedup over penalty-based methods.

Scales efficiently to networks with thousands of variables.

Abstract

Inferring the structure of directed acyclic graphs (DAGs) from data is a central challenge in causal discovery, particularly in modern high-dimensional settings where large-scale interventional data are increasingly available. While interventional data can improve identifiability, existing methods remain limited by soft acyclicity constraints, leading to optimization over invalid cyclic graphs, numerical instability, and reduced scalability. We introduce PACER (Perturbation-driven Acyclic Causal Edge Recovery), a scalable framework for causal discovery that guarantees acyclicity by construction. PACER parameterizes a distribution over DAGs through a joint model of variable permutations and edge probabilities, enabling direct optimization over valid causal structures without surrogate penalties. The framework supports a unified likelihood-based treatment of observational and interventional data, flexible conditional density models, and the incorporation of structural prior knowledge. For linear-Gaussian mechanisms, we derive closed-form expressions for the expected interventional log-likelihood and its gradients, yielding substantial computational gains. Empirically, PACER matches or exceeds state-of-the-art methods on protein signaling and large-scale genetic perturbation benchmarks, while scaling efficiently to networks with thousands of variables and achieving up to two orders of magnitude speedups over penalty-based differentiable approaches. These results demonstrate that exact and scalable causal discovery from high-dimensional perturbation data is achievable through principled search space design.