Zero-Day Attack Detection in Cloud Security: A Survey of Machine Learning and Deep Learning Approaches
L Manjula, E. Saravana Kumar · 2025
In recent times, the increasing use of cloud computing by organizations for scalability and efficiency has made it crucial to prioritize the security of their cloud environments. The use of unknown software and hardware vulnerabilities often leads to Zero-Day attacks, which exploit these weaknesses before patches can be applied. Zero-day attacks have the benefits of unrevealed vulnerabilities and it can appear while an organization is not capable to respond. Hence these attacks are particularly dangerous for the cloud-oriented services because it resulted in data breaches. In this survey, Machine Learning (ML) and Deep Learning (DL) approaches are highlighted as the primary need for attack detection mechanisms using Zero-Day in cloud security.. Hence, the survey emphasizes the importance of feature engineering, anomaly detection, and hybrid models that combine traditional security measures with advanced AI techniques to improve detection accuracy. The review also address the challenges of implementing methods such as the need for large datasets, the issue of false positives and the difficulty of interpretability. Moreover, it highlight the importance of continuous learning and adaptation in detection systems to adapt to changing threats. By providing a survey of the advanced research on Zero-Day attacks in cloud security, this survey seeks gaps from current studies and identify any breaches in existing methodologies. Ultimately, it believe in proactive, adaptive security with strong ML and dynamic DL capabilities to effectively mitigate risks of Zero-Day attacks while maintaining a more secure cloud computing environment for organizations.