Cutting-Edge Broad Learning with Flexible Privacy Mechanisms for Detecting Malicious Behaviour

Protecting sensitive user data and identifying malicious behaviour in cloud environments are significant challenges. This paper introduces a novel privacy-preserving broad learning system, designed to improve cloud security. Integrating flexible differential privacy with parallel processing, the proposed system addresses the computational efficiency and privacy protection limitations inherent in traditional deep learning models. Our approach highlights data protection through advanced privacy-preserving techniques, adding an additional layer of security. The system dynamically responds to evolving attack patterns, enhancing its accuracy in detecting potential threats. Our evaluations demonstrate that our method outperforms existing approaches in accuracy, precision, recall, and F-measure, while also demonstrating efficient response times. This marks a significant advancement in cloud computing security.