mlsanityData to 7 Oct 2026

Privacy Preserving Diffusion Models for Mixed-Type Tabular Data Generation

arXiv 2512.00638 ↗First posted 29 Nov 2025cs.LGcited by 0

On the map: Diffusion models → Robustness, privacy and adversarial aspects

Abstract (arXiv)

We introduce DP-FinDiff, a differentially private diffusion framework for synthesizing mixed-type tabular data. DP-FinDiff employs embedding-based representations for categorical features, reducing encoding overhead and scaling to high-dimensional datasets. To adapt DP-training to the diffusion process, we propose two privacy-aware training strategies: an adaptive timestep sampler that aligns updates with diffusion dynamics, and a feature-aggregated loss that mitigates clipping-induced bias. Together, these enhancements improve fidelity and downstream utility without weakening privacy guarantees. On financial and medical datasets, DP-FinDiff achieves 16-42% higher utility than DP baselines at comparable privacy levels, demonstrating its promise for safe and effective data sharing in sensitive domains.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Robustness, privacy and adversarial aspects, one of 14 directions in Diffusion models. Builds on 29 of its 43 references that are in the map. First posted on arXiv 29 Nov 2025.

Builds on

29 of 43 references.

Closest ancestors · 2

Same direction, newest first

  1. 2023Extracting Training Data from Diffusion Models30 Jan
  2. 2022Differentially Private Diffusion Models18 Oct

Other references · 27

  1. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  2. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  3. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,005
  4. 2016Deep Learning with Differential Privacycited by 2,682
  5. 2016Membership Inference Attacks against Machine Learning Modelscited by 2,221
  6. 2020Extracting Training Data from Large Language Modelscited by 2,144
  7. 2019Modeling Tabular data using Conditional GANcited by 695
  8. 2018ML-Leaks: Model and Data Independent Membership Inference Attacks and Defenses on Machine Learning Modelscited by 475
  9. 2021Opacus: User-Friendly Differential Privacy Library in PyTorchcited by 375
  10. 2022TabDDPM: Modelling Tabular Data with Diffusion Modelscited by 373
  11. 2023Mixed-Type Tabular Data Synthesis with Score-based Diffusion in Latent Spacecited by 183
  12. 2021Numerical Composition of Differential Privacycited by 174
15 more
  1. 2021CTAB-GAN: Effective Table Data Synthesizingcited by 156
  2. 2023CoDi: Co-evolving Contrastive Diffusion Models for Mixed-type Tabular Synthesiscited by 78
  3. 2020Conditional Wasserstein GAN-based Oversampling of Tabular Data for Imbalanced Learningcited by 67
  4. 2020Differentially Private Secure Multi-Party Computation for Federated Learning in Financial Applicationscited by 58
  5. 2022A Unified Framework for Quantifying Privacy Risk in Synthetic Datacited by 45
  6. 2023FinDiff: Diffusion Models for Financial Tabular Data Generationcited by 41
  7. 2023AutoDiff: combining Auto-encoder and Diffusion model for tabular data synthesizingcited by 24
  8. 2022On the Privacy Properties of GAN-generated Samplescited by 16
  9. 2019Using GANs for Sharing Networked Time Series Data: Challenges, Initial Promise, and Open Questionscited by 12
  10. 2022Federated and Privacy-Preserving Learning of Accounting Data in Financial Statement Auditscited by 8
  11. 2025Diffusion and Flow Matching Models for Tabular Data: A Surveycited by 8
  12. 2020PCAL: A Privacy-preserving Intelligent Credit Risk Modeling Framework Based on Adversarial Learningcited by 7
  13. 2021DTGAN: Differential Private Training for Tabular GANscited by 7
  14. 2023Deep Generative Models, Synthetic Tabular Data, and Differential Privacy: An Overview and Synthesiscited by 7
  15. 2023Generating tabular datasets under differential privacycited by 5

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2024Federated Diffusion Modeling with Differential Privacy for Tabular Data Synthesis
  2. 2024DP-TLDM: Differentially Private Tabular Latent Diffusion Model
  3. 2025Flow Matching for Tabular Data Synthesis
  4. 2026Differentially Private Synthetic Data via APIs 4: Tabular Data
  5. 2025Membership Inference over Diffusion-models-based Synthetic Tabular Data
  6. 2024FedTabDiff: Federated Learning of Diffusion Probabilistic Models for Synthetic Mixed-Type Tabular Data Generation
  7. 2025RAPID: Retrieval Augmented Training of Differentially Private Diffusion Models
  8. 2026FairDiffuseVQVAE: Sampling-Time Fairness in Tabular Diffusion via Conditional Refinement of Vector-Quantized Latents