Scalable Contrastive Causal Discovery under Unknown Soft Interventions

2026-03-03

AI Summary

This work addresses the challenge of causal structure identification under a single unknown soft intervention, where traditional methods are limited by Markov equivalence and cannot uniquely recover the true causal graph. The authors propose the first scalable causal discovery framework tailored to this setting, leveraging paired observational and interventional data under a shared causal structure assumption. By integrating subset-constrained causal discovery, cross-mechanism orientation rules, and PDAG aggregation within Meek closure, the method constructs a globally consistent maximal PDAG. Theoretically, it is shown to correctly transcend the constrained Ψ-equivalence class and asymptotically recover more causal edges. Experiments on synthetic data demonstrate significant improvements in structural recovery accuracy, along with strong generalization capability and computational efficiency for large-scale graphs.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Constraint Satisfaction and Optimization: Mixed Discrete/Continuous Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Bridging structured and unstructured data
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies

Abstract

Observational causal discovery is only identifiable up to the Markov equivalence class. While interventions can reduce this ambiguity, in practice interventions are often soft with multiple unknown targets. In many realistic scenarios, only a single intervention regime is observed. We propose a scalable causal discovery model for paired observational and interventional settings with shared underlying causal structure and unknown soft interventions. The model aggregates subset-level PDAGs and applies contrastive cross-regime orientation rules to construct a globally consistent maximal PDAG under Meek closure, enabling generalization to both in-distribution and out-of-distribution settings. Theoretically, we prove that our model is sound with respect to a restricted $Ψ$ equivalence class induced solely by the information available in the subset-restricted setting. We further show that the model asymptotically recovers the corresponding identifiable PDAG and can orient additional edges compared to non-contrastive subset-restricted methods. Experiments on synthetic data demonstrate improved causal structure recovery, generalization to unseen graphs with held-out causal mechanisms, and scalability to larger graphs, with ablations supporting the theoretical results.

Problem

• causal discovery
• soft interventions
• Markov equivalence class
• interventional data
• causal structure

Innovation

• contrastive causal discovery
• soft interventions
• PDAG aggregation
• Meek closure
• scalable causal inference