BlockRR: A Unified Framework of RR-type Algorithms for Label Differential Privacy

2026-02-03

AI Summary

This work addresses the lack of a unified framework for Randomized Response (RR) mechanisms in label differential privacy, which has necessitated ad hoc analyses of individual schemes. The authors propose BlockRR, the first unified RR framework that generalizes existing mechanisms through a parameterized design and incorporates a label-prior-aware weighted block strategy. BlockRR jointly optimizes privacy preservation and model utility under ε-label differential privacy guarantees. Experimental results on CIFAR-10 and its imbalanced variants demonstrate that BlockRR significantly improves both overall test accuracy and per-class average accuracy under moderate to high privacy budgets (ε ≤ 3.0), while incurring no performance degradation even under low privacy budgets.

Technology Category

• Machine Learning: Privacy
• Reasoning under Uncertainty: Stochastic Optimization
• Computer Vision: Bias, Fairness & Privacy

Application Category

• Responsible Web: Data and user privacy-enhancing technologies for the Web
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Privacy-enhancing technologies

Abstract

In this paper, we introduce BlockRR, a novel and unified randomized-response mechanism for label differential privacy. This framework generalizes existed RR-type mechanisms as special cases under specific parameter settings, which eliminates the need for separate, case-by-case analysis. Theoretically, we prove that BlockRR satisfies $\epsilon$-label DP. We also design a partition method for BlockRR based on a weight matrix derived from label prior information; the parallel composition principle ensures that the composition of two such mechanisms remains $\epsilon$-label DP. Empirically, we evaluate BlockRR on two variants of CIFAR-10 with varying degrees of class imbalance. Results show that in the high-privacy and moderate-privacy regimes ($\epsilon \leq 3.0$), our propsed method gets a better balance between test accuaracy and the average of per-class accuracy. In the low-privacy regime ($\epsilon \geq 4.0$), all methods reduce BlockRR to standard RR without additional performance loss.

Problem

• label differential privacy
• randomized response
• class imbalance
• privacy-accuracy tradeoff
• unified framework

Innovation

• BlockRR
• label differential privacy
• randomized response
• unified framework
• parallel composition