With the rapid expansion of the Internet of Things (IoT), user data has experienced exponential growth, leading to increasing concerns about the security and integrity of data stored in the cloud. Traditional schemes relying on untrusted third-party auditors suffer from both security and efficiency issues, while existing decentralized blockchain-based auditing solutions still face shortcomings in correctness and security. This paper proposes an improved blockchain-based cloud auditing scheme, with the following core contributions: Identifying critical logical contradictions in the original scheme, thereby establishing the foundation for the correctness of cloud auditing; Designing an enhanced mechanism that integrates multiple hashing with dynamic aggregate signatures, binding encrypted blocks through bilinear pairings and BLS signatures, and improving the scheme by setting parameters based on the Computational Diffie-Hellman (CDH) problem, significantly strengthening data integrity protection and anti-forgery capabilities; Introducing a random challenge mechanism and dynamic parameter adjustment strategy, effectively resisting various attacks such as forgery, tampering, and deletion, significantly improving the detection probability of malicious Cloud Service Providers (CSPs), and significantly reducing the proof generation overhead for CSPs while maintaining the same computational cost for Data Owners. Theoretical analysis and performance evaluation experiments demonstrate that the proposed scheme achieves significant improvements in both security and efficiency. Finally, the paper explores potential applications of the Enhanced Security Scheme in fields such as healthcare, drone swarms, and government office attendance systems, providing an effective approach for building secure, efficient, and decentralized cloud auditing systems.

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.