<ul><li>Differentially private (DP) machine learning often relies on the availability of public data for tasks like privacy-utility trade-off estimation, hyperparameter tuning, and pretraining.</li><li>For tabular data, the assumption of public data may not hold due to heterogeneity across domains.</li><li>To address this, the proposal is to generate surrogate public data from schema-level specifications without accessing sensitive records.</li><li>Experiments demonstrate that surrogate public tabular data can effectively replace traditional public data for tasks such as pretraining differentially private tabular classifiers.</li></ul>

Do You Really Need Public Data? Surrogate Public Data for Differential Privacy on Tabular Data

Discover more