Privacy Preserving Diffusion Models for Mixed-Type Tabular Data Generation
Fuente:
arXiv
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914174814126080 |
|---|---|
| author | Sattarov, Timur Schreyer, Marco Borth, Damian |
| author_facet | Sattarov, Timur Schreyer, Marco Borth, Damian |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_00638 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Privacy Preserving Diffusion Models for Mixed-Type Tabular Data Generation Sattarov, Timur Schreyer, Marco Borth, Damian Machine Learning 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. |
| title | Privacy Preserving Diffusion Models for Mixed-Type Tabular Data Generation |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2512.00638 |