55
30
42
30
A GAN-Based Framework for Robust DDoS Attack Detection
WGAN-GP-generated adversarial DDoS traffic augments training data, improving detection resilience against evasion attempts.
Researchers built a DDoS detection framework combining Random Forests, deep neural ensembles, and Transformer-based models trained on CICDDoS2019 with synthetic adversarial flows generated by a Wasserstein GAN with gradient penalty. Hybrid datasets of benign, malicious, and generated traffic taught models more generalizable decision boundaries. Experiments showed improved accuracy and resilience against unseen adversarial traffic, validated on real-world generated flows.
32
30
45
57
45
30
30
60
45
57
60
57
60
42
30
57
30
30
30
42
57
55
60
60
30
30
30
30
45
45
30
30
60
30
30
60