The rapid adoption of machine learning in sensitive domains,such as healthcare,finance,and government services,has heightened the need for robust,privacy-preserving techniques.Traditional machine learning approaches l...The rapid adoption of machine learning in sensitive domains,such as healthcare,finance,and government services,has heightened the need for robust,privacy-preserving techniques.Traditional machine learning approaches lack built-in privacy mechanisms,exposing sensitive data to risks,which motivates the development of Privacy-Preserving Machine Learning(PPML)methods.Despite significant advances in PPML,a comprehensive and focused exploration of Secure Multi-Party Computing(SMPC)within this context remains underdeveloped.This review aims to bridge this knowledge gap by systematically analyzing the role of SMPC in PPML,offering a structured overviewof current techniques,challenges,and future directions.Using a semi-systematicmapping studymethodology,this paper surveys recent literature spanning SMPC protocols,PPML frameworks,implementation approaches,threat models,and performance metrics.Emphasis is placed on identifying trends,technical limitations,and comparative strengths of leading SMPC-based methods.Our findings reveal thatwhile SMPCoffers strong cryptographic guarantees for privacy,challenges such as computational overhead,communication costs,and scalability persist.The paper also discusses critical vulnerabilities,practical deployment issues,and variations in protocol efficiency across use cases.展开更多
Due to the rapid advancement of information technology,data has emerged as the core resource driving decision-making and innovation across all industries.As the foundation of artificial intelligence,machine learning(M...Due to the rapid advancement of information technology,data has emerged as the core resource driving decision-making and innovation across all industries.As the foundation of artificial intelligence,machine learning(ML)has expanded its applications into intelligent recommendation systems,autonomous driving,medical diagnosis,and financial risk assessment.However,it relies on massive datasets,which contain sensitive personal information.Consequently,Privacy-Preserving Machine Learning(PPML)has become a critical research direction.To address the challenges of efficiency and accuracy in encrypted data computation within PPML,Homomorphic Encryption(HE)technology is a crucial solution,owing to its capability to facilitate computations on encrypted data.However,the integration of machine learning and homomorphic encryption technologies faces multiple challenges.Against this backdrop,this paper reviews homomorphic encryption technologies,with a focus on the advantages of the Cheon-Kim-Kim-Song(CKKS)algorithm in supporting approximate floating-point computations.This paper reviews the development of three machine learning techniques:K-nearest neighbors(KNN),K-means clustering,and face recognition-in integration with homomorphic encryption.It proposes feasible schemes for typical scenarios,summarizes limitations and future optimization directions.Additionally,it presents a systematic exploration of the integration of homomorphic encryption and machine learning from the essence of the technology,application implementation,performance trade-offs,technological convergence and future pathways to advance technological development.展开更多
文摘The rapid adoption of machine learning in sensitive domains,such as healthcare,finance,and government services,has heightened the need for robust,privacy-preserving techniques.Traditional machine learning approaches lack built-in privacy mechanisms,exposing sensitive data to risks,which motivates the development of Privacy-Preserving Machine Learning(PPML)methods.Despite significant advances in PPML,a comprehensive and focused exploration of Secure Multi-Party Computing(SMPC)within this context remains underdeveloped.This review aims to bridge this knowledge gap by systematically analyzing the role of SMPC in PPML,offering a structured overviewof current techniques,challenges,and future directions.Using a semi-systematicmapping studymethodology,this paper surveys recent literature spanning SMPC protocols,PPML frameworks,implementation approaches,threat models,and performance metrics.Emphasis is placed on identifying trends,technical limitations,and comparative strengths of leading SMPC-based methods.Our findings reveal thatwhile SMPCoffers strong cryptographic guarantees for privacy,challenges such as computational overhead,communication costs,and scalability persist.The paper also discusses critical vulnerabilities,practical deployment issues,and variations in protocol efficiency across use cases.
基金supported by the fllowing projects:Natural Science Foundation of China under Grant 62172436Self-Initiated Scientific Research Project of the Chinese People's Armed Police Force under Grant ZZKY20243129Basic Frontier Innovation Project of the Engineering University of the Chinese People's Armed Police Force under Grant WJY202421.
文摘Due to the rapid advancement of information technology,data has emerged as the core resource driving decision-making and innovation across all industries.As the foundation of artificial intelligence,machine learning(ML)has expanded its applications into intelligent recommendation systems,autonomous driving,medical diagnosis,and financial risk assessment.However,it relies on massive datasets,which contain sensitive personal information.Consequently,Privacy-Preserving Machine Learning(PPML)has become a critical research direction.To address the challenges of efficiency and accuracy in encrypted data computation within PPML,Homomorphic Encryption(HE)technology is a crucial solution,owing to its capability to facilitate computations on encrypted data.However,the integration of machine learning and homomorphic encryption technologies faces multiple challenges.Against this backdrop,this paper reviews homomorphic encryption technologies,with a focus on the advantages of the Cheon-Kim-Kim-Song(CKKS)algorithm in supporting approximate floating-point computations.This paper reviews the development of three machine learning techniques:K-nearest neighbors(KNN),K-means clustering,and face recognition-in integration with homomorphic encryption.It proposes feasible schemes for typical scenarios,summarizes limitations and future optimization directions.Additionally,it presents a systematic exploration of the integration of homomorphic encryption and machine learning from the essence of the technology,application implementation,performance trade-offs,technological convergence and future pathways to advance technological development.
