Towards Efficient Privacy-Preserving Machine Learning: A Systematic Review from Protocol, Model, and System Perspectives

2025-07-19

AI Summary

Privacy-preserving machine learning (PPML) suffers from severe efficiency bottlenecks and poor scalability due to cryptographic protocols (e.g., MPC, HE). To address this, we propose a cross-layer co-optimization paradigm integrating protocol design, model architecture, and system implementation. Our approach combines lightweight cryptographic protocols, model sparsification and quantization, hardware acceleration (FPGA/GPU), and adaptive scheduling. Through systematic qualitative and quantitative analysis, we demonstrate that inter-layer coupling optimization reduces computational overhead by orders of magnitude. We establish an open-source knowledge repository to track state-of-the-art advances, synthesize mainstream technical pathways, and characterize the fundamental trade-offs among efficiency, accuracy, and security. Furthermore, we identify promising future directions—including heterogeneous trusted execution environment integration and dynamic protocol selection—to advance practical, cloud-based privacy-preserving computation. This work provides both theoretical foundations and engineering guidelines for scalable PPML deployment.

Technology Category

• Machine Learning: Privacy
• Search and Optimization: Learning to Search
• Data Mining & Knowledge Management: Scalability, Parallel & Distributed Systems

Application Category

• Security and Privacy: Privacy-enhancing technologies
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Privacy-preserving machine learning (PPML) based on cryptographic protocols has emerged as a promising paradigm to protect user data privacy in cloud-based machine learning services. While it achieves formal privacy protection, PPML often incurs significant efficiency and scalability costs due to orders of magnitude overhead compared to the plaintext counterpart. Therefore, there has been a considerable focus on mitigating the efficiency gap for PPML. In this survey, we provide a comprehensive and systematic review of recent PPML studies with a focus on cross-level optimizations. Specifically, we categorize existing papers into protocol level, model level, and system level, and review progress at each level. We also provide qualitative and quantitative comparisons of existing works with technical insights, based on which we discuss future research directions and highlight the necessity of integrating optimizations across protocol, model, and system levels. We hope this survey can provide an overarching understanding of existing approaches and potentially inspire future breakthroughs in the PPML field. As the field is evolving fast, we also provide a public GitHub repository to continuously track the developments, which is available at https://github.com/PKU-SEC-Lab/Awesome-PPML-Papers.

Problem

• Efficiency and scalability issues in privacy-preserving machine learning
• Cross-level optimizations for PPML protocols, models, and systems
• Comprehensive review and future directions for PPML advancements

Innovation

• Cryptographic protocols for privacy protection
• Cross-level optimizations for efficiency
• Public GitHub repository for updates