PACER: Acyclic Causal Discovery from Large-Scale Interventional Data

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Torné, Ramon Viñas, Salazar, Sílvia Fàbregas, Park, Soyon, Ban, Ivo Alexander, Gadetsky, Artyom, Doikov, Nikita, Brbić, Maria
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866917497168461824
author Torné, Ramon Viñas
Salazar, Sílvia Fàbregas
Park, Soyon
Ban, Ivo Alexander
Gadetsky, Artyom
Doikov, Nikita
Brbić, Maria
author_facet Torné, Ramon Viñas
Salazar, Sílvia Fàbregas
Park, Soyon
Ban, Ivo Alexander
Gadetsky, Artyom
Doikov, Nikita
Brbić, Maria
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2605_15353
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle PACER: Acyclic Causal Discovery from Large-Scale Interventional Data
Torné, Ramon Viñas
Salazar, Sílvia Fàbregas
Park, Soyon
Ban, Ivo Alexander
Gadetsky, Artyom
Doikov, Nikita
Brbić, Maria
Machine Learning
Artificial Intelligence
Molecular Networks
Quantitative Methods
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.
title PACER: Acyclic Causal Discovery from Large-Scale Interventional Data
topic Machine Learning
Artificial Intelligence
Molecular Networks
Quantitative Methods
url https://arxiv.org/abs/2605.15353