Advancing Cybersecurity: A Machine Learning and Deep Learning-Based Intrusion Detection System A Review

Abstract

The fundamental storage and data handling habits of organizations and individuals changed due to the fact that cloud computing provides dynamic systems with speed in performance as well as competitive costs. Trusted security threats accompany the increased benefits of cloud computing that lead to viruses infecting data systems, exposing operational details through data leakages and hacking, and violating personal privacy. Legacy security solutions fail when identifying and thwarting intricate cyberattacks, which happen these days Due to enabling predictive analysis, automated intervention, and immediate assessment of threats, machine learning has emerged as an extremely powerful cloud security platform. This review examines the progress and challenges incurred by machine learning approaches when employed to secure the cloud infrastructure. The research investigates 2019 to 2025 studies to analyze prominent concepts resolved by data security and privacy automation and malware and intrusion detection and novel threats. Cyber threats were detected and defended by numerous machine learning methods effectively, such as supervised training, deep neural networks, reinforced learning, and federated learning. Recurrent neural network and convolutional neural network settings deployed within intrusion detection systems result in better performance for detecting malicious network behavior. Federated learning and privacy-preserving machine learning strategies become potential solutions to protect cloud environments and preserve user data.