DCD-PFN: A Decoupling-Aware Foundation Model for Causal Discovery
This work addresses the computational bottleneck and poor generalization of traditional causal discovery algorithms for practitioners needing reliable structure learning from tabular data.
DCD-PFN introduces a decoupling-aware foundation model for causal discovery that achieves robust zero-shot generalization by focusing on local Markov boundary identification rather than global graph reconstruction, outperforming traditional algorithms on highly non-linear and noisy systems.
Causal discovery is critical for understanding complex data-generating mechanisms, yet traditional algorithms often struggle with highly non-linear and noisy systems, or suffer from severe computational bottlenecks. Recent tabular foundation models based on Prior-Data Fitted Networks (PFNs) have demonstrated remarkable zero-shot inference capabilities, but their potential for explicit structural causal discovery remains underexplored. To bridge this gap, we propose DCD-PFN, a decoupling-aware foundation model for causal discovery. Instead of directly amortizing global graph reconstruction, DCD-PFN focuses on local causal discovery through a decoupling-based paradigm. Through pre-training on diverse synthetic Structural Causal Models (SCMs), the model learns sample-wise decoupling weights that enable Markov boundary (MB) identification. Furthermore, by leveraging parallelized local discovery, DCD-PFN efficiently reconstructs global causal graphs while remaining grounded in the theoretical foundations of decoupling-based causal discovery. Experiments demonstrate that our foundation model achieves robust zero-shot generalization.