探討合成數據的實證私隱狀況
Exploring The Empirical Privacy Landscape of Synthetic Data
研究比較了合成醫療記錄上不同資料層面及訓練層面的幹預方法,發現以具私隱意識的方式改寫合成數據,比直接微調或 DP-SGD 更能兼顧模型能力與私隱保護。相較於一般改寫,要求替換獨特私隱細節的改寫指示能進一步減少私隱洩漏;改寫模型能力較強時,下游洩漏風險通常較低。
Published on Oct 15
Authors:
,
,
,
,
Abstract
Privacy-aware rewriting of synthetic medical records improves privacy protection with less capability loss than direct fine-tuning or DP-SGD, offering a favorable trade-off for sensitive data.
Methods for training language models on sensitive data differ in how they balance capability gains against the risk of leaking private information. In this paper, we present a controlled comparison of data-level and training-level interventions for navigating the capability-privacy tradeoff on synthetic medical records with canaries to probe for leakage. Direct finetuning performs best on capability metrics but leads to the most privacy leakage, while DP-SGD reduces leakage at substantial cost to capabilities. Generating synthetic data from these records offers a more favorable balance, particularly when guided by a uniqueness-aware instruction to replace private details. Relative to naive paraphrasing, privacy-aware rewriting improves privacy protection across a variety of different rewriting models. We also find that more capable rewriting models generally reduce the risks of downstream privacy leakage, although held-out fit does not improve monotonically and increased test-time compute yields inconsistent gains. Overall, these results highlight privacy-aware rewriting as a promising direction for achieving a more favorable trade-off between capabilities and empirical privacy compared to other training-level methods, and our evaluation framework supports comparisons of future models and methods.
View arXiv page View PDF Add to collection
Get this paper in your agent:
hf papers read 2610.22659
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash
Models citing this paper 0
No model linking this paper
Cite arxiv.org/abs/2610.22659 in a model README.md to link it from this page.
Datasets citing this paper 0
No dataset linking this paper
Cite arxiv.org/abs/2610.22659 in a dataset README.md to link it from this page.
Spaces citing this paper 0
No Space linking this paper
Cite arxiv.org/abs/2610.22659 in a Space README.md to link it from this page.
Collections including this paper 0
No Collection including this paper
Add this paper to a collection to link it from this page.
來源:HuggingFace Daily Papers(社區熱門論文) · huggingface.co