Co-Evolving Zero-Day Jamming: Adaptive Attack Synthesis and Graph Attention-Based Online Detection
An online detection framework is introduced that combines a graph attention network (GAT) for temporal-spectral representation learning with Dirichlet process (DP)-means clustering and an inference-driven reinforcement learning (RL) jammer is proposed as an adversarial benchmark.
Ghilas Aissou, Re-Mi A. Chou, Taejoon Kim
· 0 citations