一句话：提出近似 span 提升，为多种差分隐私松弛提供组合语义与可证明其性质的程序逻辑。
论文针对 Rényi 差分隐私、zero-concentrated 差分隐私和 truncated concentrated 差分隐私等标准差分隐私的松弛，研究如何组合式地证明两个输出概率分布之间的统计散度界。作者提出近似 span 提升（approximate span lifting），将此前用于标准差分隐私的近似关系提升推广到更一般的散度类和连续分布，从而为这些隐私松弛提供统一的组合语义。基于该构造，论文进一步设计了程序逻辑，可证明差分隐私松弛及其他统计散度性质。关键结果是建立了可组合的语义与逻辑框架，将多种隐私/散度保证纳入同一推理体系。
We develop new abstractions for reasoning about three relaxations of differential privacy: Rényi differential privacy, zero-concentrated differential privacy, and truncated concentrated differential privacy, which express bounds on statistical divergences between two output probability distributions. In order to reason about such properties compositionally, we introduce approximate span-lifting, a novel construction extending the approximate relational lifting approaches previously developed for standard differential privacy to a more general class of divergences, and also to continuous distributions. As an application, we develop a program logic based on approximate span-liftings capable of proving relaxations of differential privacy and other statistical divergence properties.
DOI 原文 ·
@inproceedings{paperbot5006,
title = {Approximate Span Liftings: Compositional Semantics for Relaxations of Differential Privacy},
author = {Tetsuya Sato and Gilles Barthe and Marco Gaboardi and Justin Hsu and Shin-ya Katsumata},
booktitle = {LICS},
year = {2019},
doi = {10.1109/lics.2019.8785668}
}