Multi-class Item Mining under Local Differential Privacy

2025/04/18 by Mao, Yulian, Ye, Qingqing, Du, Rong +3
#Cryptography and Security (cs.CR) #FOS: Computer and information sciences

paper · doi:10.48550/arxiv.2504.13526

Abstract

Item mining, a fundamental task for collecting statistical data from users, has raised increasing privacy concerns. To address these concerns, local differential privacy (LDP) was proposed as a privacy-preserving technique. Existing LDP item mining mechanisms primarily concentrate on global statistics, i.e., those from the entire dataset. Nevertheless, they fall short of user-tailored tasks such as personalized recommendations, whereas classwise statistics can improve task accuracy with fine-grained information. Meanwhile, the introduction of class labels brings new challenges. Label perturbation may result in invalid items for aggregation. To this end, we propose frameworks for multi-class item mining, along with two mechanisms: validity perturbation to reduce the impact of invalid data, and correlated perturbation to preserve the relationship between labels and items. We also apply these optimized methods to two multi-class item mining queries: frequency estimation and top-k item mining. Through theoretical analysis and extensive experiments, we verify the effectiveness and superiority of these methods.

Citations

• Real-Time Trajectory Synthesis with Local Differential Privacy
• Local Differentially Private Heavy Hitter Detection in Data Streams with Bounded Memory
• Machine Learning-Based Heart Disease Diagnosis: A Systematic Literature Review
• Machine learning-based heart disease diagnosis: A systematic literature review
• Antipodes of Label Differential Privacy: PATE and ALIBI
• Deep Learning with Label Differential Privacy
• Estimating Numerical Distributions under Local Differential Privacy
• Data Poisoning Attacks to Local Differential Privacy Protocols
• Collecting and Analyzing Multidimensional Data with Local Differential Privacy
• Locally Differentially Private Heavy Hitter Identification
• RAPPOR: Randomized Aggregatable Privacy-Preserving Ordinal Response
• What Can We Learn Privately?
• Differential Privacy
• Cumulated gain-based evaluation of IR techniques
• Machine learning for medical diagnosis: history, state of the art and perspective
• Word association norms, mutual information, and lexicography

Related

• Production of Categorical Data Verifying Differential Privacy: Conception and Applications to Machine Learning
• Collecting and Analyzing Multidimensional Data with Local Differential Privacy
• A Framework for High-Accuracy Privacy-Preserving Mining
• Local Privacy-preserving Mechanisms and Applications in Machine Learning
• Element Level Differential Privacy: The Right Granularity of Privacy