dailyai.report

A new framework combines multi-layer perceptron, SMOTE, and non-IID federated learning to detect network intrusions. The system integrates explainable AI to clarify why specific traffic is flagged as malicious. This approach addresses data imbalance and privacy in distributed environments. Security practitioners can now better audit automated threat detections without compromising raw local data.