Tianhang Zheng (University of Missouri-Kansas City), Baochun Li (University of Toronto)

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.

View More Papers

Eavesdropping on Controller Acoustic Emanation for Keystroke Inference Attack...

Shiqing Luo (George Mason University), Anh Nguyen (George Mason University), Hafsa Farooq (Georgia State University), Kun Sun (George Mason University), Zhisheng Yan (George Mason University)

Read More

AnonPSI: An Anonymity Assessment Framework for PSI

Bo Jiang (TikTok Inc.), Jian Du (TikTok Inc.), Qiang Yan (TikTok Inc.)

Read More

Parrot-Trained Adversarial Examples: Pushing the Practicality of Black-Box Audio...

Rui Duan (University of South Florida), Zhe Qu (Central South University), Leah Ding (American University), Yao Liu (University of South Florida), Zhuo Lu (University of South Florida)

Read More

Bernoulli Honeywords

Ke Coby Wang (Duke University), Michael K. Reiter (Duke University)

Read More