A COMPREHENSIVE SURVEY ON AI-ENHANCED CYBER THREAT DETECTION: MODELS, DATASETS, AND METRICS,

The escalation of cyber threats in both frequency and complexity has posed significant challenges to traditional cybersecurity mechanisms. In response to these evolving threats, Use of Artificial Intelligence in Thwarting Cyber Threats The use of AI is increasingly becoming essential to fortify detection of threats, augmenting preventive actions and bettering responses to incidents. Human in the loop AI Recent AI breakthroughs have spawned a new generation of models, which can process large data sets, can detect anomalous patterns and predict potential attacks with unprecedented accuracy and speed. This survey provides a comprehensive review of the recent research studies that propose AI-based techniques for various cyber threat detection scenarios. The paper categorizes these approaches based on their underlying models, such as transfer learning, ensemble methods, graph neural networks, and GAN. Additionally, provide a comparative analysis of their methodologies, performance metrics, addressed challenges, and future work. The purpose of this paper is to assist the practitioners and researchers in choosing and enhancing any AI based solutions relevant to cybersecurity.