This project studies how adversarial attacks can fool ML-based Network Intrusion Detection Systems (NIDS), and how well adversarial retraining can recover robustness. The attacks are constrained to be ...