AI-Powered Autonomous Anomaly Detection in IAM: A Quantum-Resistant Deep Learning Framework with Real-Time Adaptive Risk Management and Federated Learning

Abstract

This study introduces the Quantum-Resistant Federated Anomaly Detection (QRFAD) framework, a novel approach for addressing the challenges of anomaly detection in Identity and Access Management (IAM) systems in the face of evolving cyber threats and emerging quantum computing risks. By combining federated learning, quantum-resistant cryptography, and Agentic AI, QRFAD overcomes the limitations of traditional anomaly detection models, such as LSTM and SVM. The framework ensures a scalable, secure, and privacy-preserving solution that enables multiple organizations to collaboratively train models without sharing sensitive data. Key performance metrics, including precision, recall, F1-score, latency, and scalability, were evaluated on the ICS-Flow dataset, demonstrating that QRFAD significantly outperformed the LSTM and SVM models. Specifically, QRFAD achieved a precision of 0.94, recall of 0.91, and F1-score of 0.92, all higher than the LSTM (0.85, 0.83, 0.84) and SVM models (0.78, 0.75, 0.76). Moreover, QRFAD reduced the latency by 35.71\% (75 ms vs. 120 ms) and improved the model update overhead by 50\% (1.2s vs. 2.5s). This work addresses key gaps in the literature related to scalability, quantum threats, and autonomous decision-making, providing an adaptive and secure IAM solution capable of evolving with emerging cyber threats.