HeteroFedSyn: Differentially Private Tabular Data Synthesis for Heterogeneous Federated Settings

2026-03-09

AI Summary

This work addresses the challenge of distribution heterogeneity in horizontally federated settings, which often leads to high bias or excessive noise in existing differentially private tabular data synthesis methods. The authors propose a second-order marginal-based synthetic data framework tailored for horizontal federated learning, which innovatively integrates an L2 dependence measure with random projections to efficiently capture feature correlations. They design an unbiased estimator under multiplicative noise and introduce an adaptive marginal selection strategy that dynamically updates dependence scores. Under strict differential privacy guarantees, the method significantly improves synthetic data utility, achieving performance on par with centralized approaches in range queries, Wasserstein distance, and downstream machine learning tasks, thereby effectively balancing privacy preservation and data fidelity.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Security and Privacy: Data transparency and provenance
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing

Abstract

Traditional Differential Privacy (DP) mechanisms are typically tailored to specific analysis tasks, which limits the reusability of protected data. DP tabular data synthesis overcomes this by generating synthetic datasets that can be shared for arbitrary downstream tasks. However, existing synthesis methods predominantly assume centralized or local settings and overlook the more practical horizontal federated scenario. Naively synthesizing data locally or perturbing individual records either produces biased mixtures or introduces excessive noise, especially under heterogeneous data distributions across participants. We propose HeteroFedSyn, the first DP tabular data synthesis framework designed specifically for the horizontal federated setting. Built upon the PrivSyn paradigm of 2-way marginal-based synthesis, HeteroFedSyn introduces three key innovations for distributed marginal selection: (i) an L2-based dependency metric with random projection for noise-efficient correlation measurement, (ii) an unbiased estimator to correct multiplicative noise, and (iii) an adaptive selection strategy that dynamically updates dependency scores to avoid redundancy. Extensive experiments on range queries, Wasserstein fidelity, and machine learning tasks show that, despite the increased noise inherent to federated execution, HeteroFedSyn achieves utility comparable to centralized synthesis. Our code is open-sourced via the link.

Problem

• Differential Privacy
• Tabular Data Synthesis
• Federated Learning
• Data Heterogeneity
• Horizontal Federated Setting

Innovation

• Differential Privacy
• Federated Learning
• Tabular Data Synthesis
• Marginal Selection
• Heterogeneous Data