Vorapong Suppakitpaisarn

Kevin Pfisterer, Quentin Hillebrand, Vorapong Suppakitpaisarn

We propose an algorithm for counting below-threshold triangles in weighted graphs under local weight differential privacy. While prior work has largely focused on unweighted graphs, edge weights are intrinsic to many real-world networks. We consider the setting in which the graph topology is publicly known and privacy is required only for the contribution of an individual to incident edge weights, capturing practical scenarios such as road and telecommunication networks. Our method uses two rounds of communication. In the first round, each node releases privatized information about its incident edge weights under local weight differential privacy. In the second round, nodes locally count below-threshold triangles using this privatized information; we introduce both biased and unbiased variants of the estimator. We further develop two refinements: (i) a pre-computation step that reduces covariance and thus lowers expected error, and (ii) an efficient procedure for computing smooth sensitivity, which substantially reduces running time relative to a straightforward implementation. Finally, we present experimental results that quantify the trade-offs between the biased and unbiased variants and demonstrate the effectiveness of the proposed improvements.

Jing Bi, Vorapong Suppakitpaisarn

This study explores the robustness of learning by symmetric loss on private data. Specifically, we leverage exponential mechanism (EM) on private labels. First, we theoretically re-discussed properties of EM when it is used for private learning with symmetric loss. Then, we propose numerical guidance of privacy budgets corresponding to different data scales and utility guarantees. Further, we conducted experiments on the CIFAR-10 dataset to present the traits of symmetric loss. Since EM is a more generic differential privacy (DP) technique, it being robust has the potential for it to be generalized, and to make other DP techniques more robust.

Quentin Hillebrand, Vorapong Suppakitpaisarn, Tetsuo Shibuya

We suggest the use of hash functions to cut down the communication costs when counting subgraphs under edge local differential privacy. While various algorithms exist for computing graph statistics, including the count of subgraphs, under the edge local differential privacy, many suffer with high communication costs, making them less efficient for large graphs. Though data compression is a typical approach in differential privacy, its application in local differential privacy requires a form of compression that every node can reproduce. In our study, we introduce linear congruence hashing. With a sampling rate of $s$, our method can cut communication costs by a factor of $s^2$, albeit at the cost of increasing variance in the published graph statistic by a factor of $s$. The experimental results indicate that, when matched for communication costs, our method achieves a reduction in the $\ell_2$-error for triangle counts by up to 1000 times compared to the performance of leading algorithms.