Connecting the Dots: Generating Realistic Tabular Data with Structural Causal Models

ICLR 2026 Conference Submission17680 Authors

Keywords: Tabular data generation, Privacy-Preserving, Synthetic Data, Realism, Data Synthesis

TL;DR: TabSCM is a fast, interpretable tabular-data generator that preserves causal structure, outperforms existing models in statistical fidelity, utility, and privacy, and supports counterfactual reasoning and out-of-distribution generation.

Abstract: Most tabular-data generators match marginal statistics yet ignore causal structure, leading downstream models to learn spurious or unfair patterns. We present TabSCM, a mixed-type generator that preserves those causal dependencies. Starting from a Completed Partially Directed Acyclic Graph (CPDAG) found by any discovery algorithm, TabSCM (i) orients edges to a DAG, (ii) fits root-node marginals with KDE or categorical frequencies, and (iii) learns topologically ordered structural assignments: conditional diffusion models for continuous children and gradient-boosted trees for categorical ones. Ancestral sampling yields semantically valid records and enables exact counterfactual queries. On seven public datasets, encompassing healthcare, finance, housing, environment, TabSCM matches or surpasses state-of-the-art GAN, diffusion, and LLM baselines in statistical fidelity, downstream utility, and privacy risk, while also cutting rule-violation rates and providing causally meaningful and robust counterfactual interventions. Because generation is decomposed into explicit equations, it runs up to 583$\times$ faster than diffusion-only models and exposes interpretable knobs for fairness auditing and policy simulation, making TabSCM a practical choice for realism, explainability, and causal soundness.

Primary Area: generative models

Submission Number: 17680