MetaCaDI: A Meta-Learning Framework for Scalable Causal Discovery with Unknown Interventions

2025-10-25

AI Summary

Causal graph discovery under high data costs and unknown interventions remains challenging. Method: This paper proposes the first framework unifying causal structure learning and intervention target identification as a meta-learning task. It replaces bilevel optimization with an analytical adaptive mechanism, integrating Bayesian modeling and closed-form parameter updates to enable cross-experiment sharing of causal structure estimation and intervention localization. Contributions/Results: We establish the first meta-learning paradigm jointly addressing causal discovery and unknown intervention identification; avoid implicit gradient differentiation, thereby significantly improving training stability and few-shot generalization. On synthetic and real-world gene expression datasets, our method achieves high-accuracy causal graph recovery and intervention node identification using only 10 samples—outperforming state-of-the-art approaches substantially.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Search and Optimization: Metareasoning and Metaheuristics

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Uncovering the underlying causal mechanisms of complex real-world systems remains a significant challenge, as these systems often entail high data collection costs and involve unknown interventions. We introduce MetaCaDI, the first framework to cast the joint discovery of a causal graph and unknown interventions as a meta-learning problem. MetaCaDI is a Bayesian framework that learns a shared causal graph structure across multiple experiments and is optimized to rapidly adapt to new, few-shot intervention target prediction tasks. A key innovation is our model's analytical adaptation, which uses a closed-form solution to bypass expensive and potentially unstable gradient-based bilevel optimization. Extensive experiments on synthetic and complex gene expression data demonstrate that MetaCaDI significantly outperforms state-of-the-art methods. It excels at both causal graph recovery and identifying intervention targets from as few as 10 data instances, proving its robustness in data-scarce scenarios.

Problem

• Develops scalable causal discovery framework for unknown interventions
• Learns shared causal graph across multiple experimental datasets
• Enables few-shot intervention target prediction with limited data

Innovation

• Meta-learning framework for causal discovery
• Bayesian model with analytical adaptation
• Closed-form solution bypasses gradient optimization