Accuracy anomaly detection in user-level network traffic is crucial for network security. Compared with existing models that passively detect specific anomaly classes with large labeled training ...