Beyond Distributional Fidelity: Causal-Penalized Diffusion for Synthetic Tabular Data

Key points

• Synthetic tabular generators are commonly optimized for distributional fidelity, but statistical similarity alone does not guarantee preservation of causal effects.1
• Causal Fidelity is defined with respect to a target estimand as the discrepancy between inferential distributions obtained from real and synthetic data, and theoretical results show that high statistical fidelity does not generally imply high causal fidelity.1
• We then propose a causal-fidelity-aware training framework which adds a causal discrepancy penalty to the generative objective.1
• We further establish conditions under which causal regularization improves expected causal fidelity.1

Sources (1)

• [1]Beyond Distributional Fidelity: Causal-Penalized Diffusion for Synthetic Tabular Data
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 8, 07:36 AM

  In this paper, we study whether causal fidelity can be improved directly within a fully generative tabular model.

  Synthetic tabular generators are commonly optimized for distributional fidelity, but statistical similarity alone does not guarantee preservation of causal effects.

Extractive summary: sentences quoted from the sources.