Advanced Persistent Threat Identification in Cloud Infrastructures Using Tensor-Based Machine Learning Approaches
S Sreejith Sreekandan Nair, J Muralidharan · 2025
Advanced Persistent Threats (APTs) pose a significant challenge to cloud infrastructures due to their stealthy, multi-stage attack strategies. This chapter explores the role of tensor-based machine learning approaches in identifying APTs by leveraging the multi-dimensional nature of cloud security data. Traditional machine learning models often struggle to analyze large-scale, complex data generated in cloud environments. Tensor-based techniques, such as decomposition and factorization, provide effective methods for extracting hidden patterns, anomalies, and APT indicators across temporal, spatial, and user behavior dimensions. The chapter also addresses critical challenges, including latency, scalability, and real-time implementation of tensor models in dynamic cloud infrastructures. By comparing tensor-based methods with traditional approaches, the advantages in handling high-dimensional data are demonstrated. Finally, optimization strategies and distributed frameworks are discussed to enhance real-time APT detection. This work contributes to advancing cloud security systems through efficient, scalable, and robust tensor-based methodologies.