Querying Kernel Methods Suffices for Reconstructing their Training Data
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_ | 1866915304009891840 |
|---|---|
| author | Barzilai, Daniel Margalit, Yuval Gronich, Eitan Yehudai, Gilad Galun, Meirav Basri, Ronen |
| author_facet | Barzilai, Daniel Margalit, Yuval Gronich, Eitan Yehudai, Gilad Galun, Meirav Basri, Ronen |
| contents | Over-parameterized models have raised concerns about their potential to memorize training data, even when achieving strong generalization. The privacy implications of such memorization are generally unclear, particularly in scenarios where only model outputs are accessible. We study this question in the context of kernel methods, and demonstrate both empirically and theoretically that querying kernel models at various points suffices to reconstruct their training data, even without access to model parameters. Our results hold for a range of kernel methods, including kernel regression, support vector machines, and kernel density estimation. Our hope is that this work can illuminate potential privacy concerns for such models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_19019 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Querying Kernel Methods Suffices for Reconstructing their Training Data Barzilai, Daniel Margalit, Yuval Gronich, Eitan Yehudai, Gilad Galun, Meirav Basri, Ronen Machine Learning Over-parameterized models have raised concerns about their potential to memorize training data, even when achieving strong generalization. The privacy implications of such memorization are generally unclear, particularly in scenarios where only model outputs are accessible. We study this question in the context of kernel methods, and demonstrate both empirically and theoretically that querying kernel models at various points suffices to reconstruct their training data, even without access to model parameters. Our results hold for a range of kernel methods, including kernel regression, support vector machines, and kernel density estimation. Our hope is that this work can illuminate potential privacy concerns for such models. |
| title | Querying Kernel Methods Suffices for Reconstructing their Training Data |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2505.19019 |