Differentially Private Dataset Condensation
Tianhang Zheng, Baochun Li · 2024
Recent work in ICML'22 established a connection between dataset condensation (DC) and differential privacy (DP), which is unfortunately problematic.To correctly connect DC and DP, we propose two differentially private dataset condensation (DPDC) algorithms-LDPDC and NDPDC.LDPDC is a linear DC algorithm that can be executed on a low-end Central Processing Unit (CPU), while NDPDC is a nonlinear DC algorithm that leverages neural networks to extract and match the latent representations between real and synthetic data.Through extensive evaluations, we demonstrate that LDPDC has comparable performance to recent DP generative methods despite its simplicity.NDPDC provides acceptable DP guarantees with a mild utility loss, compared to distribution matching (DM).Additionally, NDPDC allows a flexible trade-off between the synthetic data utility and DP budget.We conduct extensive experiments to evaluate our DPDC algorithms on multiple datasets, including MNIST, FMNIST, CIFAR10, and CelebA.We mainly compare our DPDC algorithms with a non-private dataset condensation method, i.e., distribution matching [8], and two recent differentially private