Few-Shot Network Intrusion Detection Using Online Triplet Mining

In three linesNetwork intrusion detection system using triplet networks with online triplet mining and KNN classifier. Few-shot approach detects attacks with as few as 10 malicious samples per class, outperforming standard supervised methods on small datasets and anomaly detection models (high false-positive rates).

Summary generated by Claude — human-verified