Md. Iqbal Hossan

Decentralized model training can enhance cybersecurity while preserving privacy, reducing communication overhead, and improving threat detection in distributed systems.

Traditional cybersecurity methods are failing against evolving threats, but this new framework combines XGBoost with SHAP to deliver interpretable and reliable intrusion detection for critical infrastructure.

A hybrid CNN-LSTM framework achieves superior accuracy in detecting complex cyber threats, outperforming traditional methods in real-time scenarios.