Understanding the Impact of Data Domain Extraction on Synthetic Data Privacy
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866909578018422784 |
|---|---|
| author | Ganev, Georgi Annamalai, Meenatchi Sundaram Muthu Selva Mahiou, Sofiane De Cristofaro, Emiliano |
| author_facet | Ganev, Georgi Annamalai, Meenatchi Sundaram Muthu Selva Mahiou, Sofiane De Cristofaro, Emiliano |
| contents | Privacy attacks, particularly membership inference attacks (MIAs), are widely used to assess the privacy of generative models for tabular synthetic data, including those with Differential Privacy (DP) guarantees. These attacks often exploit outliers, which are especially vulnerable due to their position at the boundaries of the data domain (e.g., at the minimum and maximum values). However, the role of data domain extraction in generative models and its impact on privacy attacks have been overlooked. In this paper, we examine three strategies for defining the data domain: assuming it is externally provided (ideally from public data), extracting it directly from the input data, and extracting it with DP mechanisms. While common in popular implementations and libraries, we show that the second approach breaks end-to-end DP guarantees and leaves models vulnerable. While using a provided domain (if representative) is preferable, extracting it with DP can also defend against popular MIAs, even at high privacy budgets. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_08254 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Understanding the Impact of Data Domain Extraction on Synthetic Data Privacy Ganev, Georgi Annamalai, Meenatchi Sundaram Muthu Selva Mahiou, Sofiane De Cristofaro, Emiliano Cryptography and Security Machine Learning Privacy attacks, particularly membership inference attacks (MIAs), are widely used to assess the privacy of generative models for tabular synthetic data, including those with Differential Privacy (DP) guarantees. These attacks often exploit outliers, which are especially vulnerable due to their position at the boundaries of the data domain (e.g., at the minimum and maximum values). However, the role of data domain extraction in generative models and its impact on privacy attacks have been overlooked. In this paper, we examine three strategies for defining the data domain: assuming it is externally provided (ideally from public data), extracting it directly from the input data, and extracting it with DP mechanisms. While common in popular implementations and libraries, we show that the second approach breaks end-to-end DP guarantees and leaves models vulnerable. While using a provided domain (if representative) is preferable, extracting it with DP can also defend against popular MIAs, even at high privacy budgets. |
| title | Understanding the Impact of Data Domain Extraction on Synthetic Data Privacy |
| topic | Cryptography and Security Machine Learning |
| url | https://arxiv.org/abs/2504.08254 |