Activities per year
Machine Learning (ML) has proven to be effective in many application domains. However, ML methods can be vulnerable to adversarial attacks, in which an attacker tries to fool the classification/prediction mechanism by crafting the input data. In the case of ML-based Network Intrusion Detection Systems (NIDSs), the attacker might use their knowledge of the intrusion detection logic to generate malicious traffic that remains undetected. One way to solve this issue is to adopt adversarial training, in which the training set is augmented with adversarial traffic samples. This paper presents an adversarial training approach called GADoT, which leverages a Generative Adversarial Network (GAN) to generate adversarial DDoS samples for training. We show that a state-of-the-art NIDS with high accuracy on popular datasets can experience more than 60% undetected malicious flows under adversarial attacks. We then demonstrate how this score drops to 1.8% or less after adversarial training using GADoT.
|Title of host publication||Ninth IEEE Conference on Communications and Network Security (IEEE CNS 2021): Proceedings|
|Publisher||Institute of Electrical and Electronics Engineers Inc.|
|Publication status||Published - 10 Feb 2022|
|Event||IEEE Conference on Communications and Network Security - Virtual|
Duration: 04 Oct 2021 → 06 Oct 2021
|Conference||IEEE Conference on Communications and Network Security|
|Period||04/10/2021 → 06/10/2021|
FingerprintDive into the research topics of 'GaDoT: GAN-based Adversarial Training for Robust DDoS Attack Detection'. Together they form a unique fingerprint.
99.99% accurate - What's the problem?
Sandra Scott-Hayward (Keynote speaker)02 Sep 2022
Activity: Talk or presentation types › Invited or keynote talk at national or international conference
Cyber AI - Panel
Sandra Scott-Hayward (Speaker)17 May 2022
Activity: Talk or presentation types › Oral presentation
- 1 Article
Securing AI-based Security SystemsScott-Hayward, S., 01 Jun 2022, GCSP Strategic Security Analysis, 25.
Research output: Contribution to specialist publication › ArticleOpen AccessFile