Book Title: Multidisciplinary Research in the Age of Artificial Intelligence
Chief Editors: Dr. Jagdish Kumar Sahu, and Dr. Byram Anand
Associate Editors: Dr. J. Jayamani, and Albeen Josebh Ahmed
Co-Editors: Dr. Sunakshi Verma, and Mr. Pankaj Bharat Devre
ISBN: 978-93-7183-018-8
Chapter: 10
DOI: https://doi.org/10.59646/781/10
Author: Dr. Reena Saini
Abstract
Cloud computing has evolved at an unprecedented pace, making today’s digital infrastructure much more complex, and therefore more challenging when it comes to identifying and reacting to complex cyber anomalies. This research paper highlights the application of hybrid neural-rule systems in adaptive anomaly detection for enhancing the security, accuracy, and real-time threat identification in cloud environments. This study has a mixed-method research approach, an approach that integrates quantitative analysis of the anomaly detection performance metrics and qualitative data obtained from the cybersecurity professionals and cloud system administrators. The results show that the traditional rule-based detection method offers good interpretability and reliability in detecting known attacks whereas the neural network-based detection method exhibits better performance in detecting previously unknown attacks and complex attack patterns. But data-driven models have issues of explainability, false positive and adaptability with dynamic cloud infrastructures. By incorporating rule-based reasoning into the neural network framework, a balanced detection system can be achieved, bringing together both intelligent pattern recognition and decision-making processes that are clearly defined and understandable. Moreover, hybrid systems are more adaptive in responding to security threats because they learn and adapt their security responses based on the changing behaviors and constraints on the network. The study also identifies the critical challenges of implementation such as computational complexity, data privacy, model training need, and scalability in the cloud for large-scale applications. In this paper, we introduce some key strategies to create effective hybrid anomaly detection systems using cutting-edge machine learning algorithms, explainable AI methods, and adaptive rule management systems. This research offers insights into the capabilities of a hybrid neural-rule architecture and will help drive intelligent cloud security systems that will ensure accuracy, resilience, and proactive threat management.
Keywords: Hybrid neural-rule systems, Cloud computing security, Adaptive anomaly detection, Artificial intelligence, Machine learning, Cybersecurity, Intrusion detection, Explainable AI, Cloud environments, Threat detection.