S. Ramya

AI-Powered Threat Detection Systems in Cybersecurity

The complexity of emerging cyber threats renders the need for intelligent, adaptive and real-time responses essential. In this paper, an explainable deep learning assisted federated learning and adversarial robustness-based AI-enabled threat detection scheme for cybersecurity is introduced. The framework also breaks the limitations of traditional models with continuous learning, cross-platform fusion, proactive exception detection. By using lightweight neural architectures and fusing threat information, the system achieves scale and low latency. Experiments demonstrate a 96.8% attack detection rate, and 42% decrease in false alarm rate outperforming the state-of-the-art.