RSSDL-IDS: A Real-Time Self-Supervised Deep Learning Framework for Detecting Known and Zero-Day Cyberattacks with Explainable Aler

The major research contribution of RSSDL-IDS is a unified seven-stage intrusion detection framework that addresses several limitations of conventional IDS simultaneously: limited zero-day detection, dependence on labelled data, fixed anomaly thresholds, adversarial vulnerability, concept drift, and lack of interpretable alerts. It integrates these capabilities into one pipeline and empirically evaluates their individual and combined effects through ablation experiments.