Analysis of Data Privacy Breaches Using Deep Learning in Cloud Environments: A Review
Abdulqawi Mohammed Almosti, M. M. Hafizur Rahman · Electronics · 2025
Despite the advantages of using cloud computing, data breaches and security challenges remain, especially when dealing with sensitive information. The integration of deep learning (DL) techniques in a cloud environment ensures privacy preservation. This review paper analyzes 38 papers published from 2020 to 2025, focusing on privacy-preserving techniques in DL for cloud environments. Combining different privacy preservation technologies with DL results in improved utility for privacy protection and better security against data breaches than using individual applications such as differential privacy, homomorphic encryption, or federated learning. Further, a discussion is provided on the technical limitations when applying DL with various privacy preservation techniques, which include large communication overhead, lower model accuracy, and high computational cost. Additionally, this review paper presents the latest research in a comprehensive manner and provides directions for future research necessary to develop privacy-preserving DL models.