A Joint Exponential Mechanism For Differentially Private Top-$k$

Gillenwater, Jennifer; Joseph, Matthew; Medina, Andrés Muñoz; Ribero, Mónica

doi:10.48550/arxiv.2201.12333

Cited by 1 publication

(1 citation statement)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Motivation of an adaptive k. Recent work [Carvalho et al, 2020, Durfee and Rogers, 2019, Gillenwater et al, 2022] make use of structures in the top-k counts, showing that large gaps improve the performance of the private top-k mechanisms. This leads to one natural question -can't we just set a k, such that there exists a large gap between the k and the (k + 1)th vote?…”

Section: Preliminarymentioning

confidence: 99%

Adaptive Private-K-Selection with Adaptive K and Application to Multi-label PATE

Zhu¹,

Wang²

2022

Preprint

View full text Add to dashboard Cite

We provide an end-to-end Renyi DP based-framework for differentially private top-k selection. Unlike previous approaches, which require a data-independent choice on k, we propose to privately release a data-dependent choice of k such that the gap between k-th and the (k + 1)st "quality" is large. This is achieved by a novel application of the Report-Noisy-Max. Not only does this eliminate one hyperparameter, the adaptive choice of k also certifies the stability of the top-k indices in the unordered set so we can release them using a variant of propose-test-release (PTR) without adding noise. We show that our construction improves the privacy-utility trade-offs compared to the previous top-k selection algorithms theoretically and empirically. Additionally, we apply our algorithm to "Private Aggregation of Teacher Ensembles (PATE)" in multi-label classification tasks with a large number of labels and show that it leads to significant performance gains.

show abstract

Section: Preliminarymentioning

confidence: 99%