Yassine Mekdad

Yassine Mekdad

Auto-generated profile, based on publicly available information

Yassine Mekdad

Cyber-Physical Systems Security Lab, School of Computing and Information Sciences, Florida International University, Miami, FL, USA

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
10 February 2023 in Computer Networks
Verified
Journal Article
Closed access
YMYassine Mekdad
Yassine Mekdad
AAAhmet Aris
Ahmet Aris
LBLeonardo Babun
Leonardo Babun
AFAbdeslam El Fergougui
Abdeslam El Fergougui
MCMauro Conti
Mauro Conti
AUA. Selcuk Uluagac
A. Selcuk Uluagac

1 April 2022 in IEEE Transactions on Network and Service Management
Verified
Journal Article
Closed access
ENEhsan Nowroozi
Ehsan Nowroozi
YMYassine Mekdad
Yassine Mekdad
MBMohammad Hajian Berenjestanaki
Mohammad Hajian Berenjestanaki
MCMauro Conti
Mauro Conti
AFAbdeslam El Fergougui
Abdeslam El Fergougui
Convolutional Neural Networks (CNNs) models are one of the most frequently used deep learning networks, and extensively used in both academia and industry. Recent studies demonstrated that adversarial attacks against such models can maintain their effectiveness even when used on models other than the one targeted by the attacker. This major property is known as transferability, and makes CNNs ill-suited for security applications. In this paper, we provide the first comprehensive study which assesses the robustness of CNN-based models for computer networks against adversarial transferability. Furthermore, we investigate whether the transferability property issue holds in computer networks applications. In our experiments, we first consider five different attacks: the Iterative Fast Gradient Method (I-FGSM), the Jacobian-based Saliency Map (JSMA), the Limited-memory Broyden Fletcher Goldfarb Shanno BFGS (L-BFGS), the Projected Gradient Descent (PGD), and the DeepFool attack. Then, we perform these attacks against three well-known datasets: the Network-based Detection of IoT (N-BaIoT) dataset, the Domain Generating Algorithms (DGA) dataset, and the RIPE Atlas dataset. Our experimental results show clearly that the transferability happens in specific use cases for the I-FGSM, the JSMA, and the LBFGS attack. In such scenarios, the attack success rate on the target network range from 63.00% to 100%. Finally, we suggest two shielding strategies to hinder the attack transferability, by considering the Most Powerful Attacks (MPAs), and the mismatch LSTM architecture.

13 November 2023 in IEEE Transactions on Services Computing
Verified
Journal Article
Closed access
ENEhsan Nowroozi
Ehsan Nowroozi
MMMohammadreza Mohammadi
Mohammadreza Mohammadi
PGPargol Golmohammadi
Pargol Golmohammadi
YMYassine Mekdad
Yassine Mekdad
MCMauro Conti
Mauro Conti
SUSelcuk Uluagac
Selcuk Uluagac
In the past decades, the rise of artificial intelligence has given us the capabilities to solve the most challenging problems in our day-to-day lives, such as cancer prediction and autonomous navigation. However, these applications might not be reliable if not secured against adversarial attacks. In addition, recent works demonstrated that some adversarial examples are transferable across different models. Therefore, it is crucial to avoid such transferability via robust models that resist adversarial manipulations. In this paper, we propose a feature randomization-based approach that resists eight adversarial attacks targeting deep learning models in the testing phase. Our novel approach consists of changing the training strategy in the target network classifier and selecting random feature samples. We consider the attacker with a Limited-Knowledge and Semi-Knowledge conditions to undertake the most prevalent types of adversarial attacks. We evaluate the robustness of our approach using the well-known UNSW-NB15 datasets that include realistic and synthetic attacks. Afterward, we demonstrate that our strategy outperforms the existing state-of-the-art approach, such as the Most Powerful Attack, which consists of fine-tuning the network model against specific adversarial attacks. Further, we demonstrate the practicality of our approach using the VIPPrint dataset through a comprehensive set of experiments. Finally, our experimental results show that our methodology can secure the target network and resists adversarial attack transferability by over 60%.

3 August 2023 in Security and Privacy
Verified
Journal Article
Closed access
YMYassine Mekdad
Yassine Mekdad
AAAhmet Aris
Ahmet Aris
AAAbbas Acar
Abbas Acar
MCMauro Conti
Mauro Conti
RLRiccardo Lazzeretti
Riccardo Lazzeretti
SUSelcuk Uluagac
Selcuk Uluagac

8 February 2022
Verified
Book Chapter
Closed access
PDF file
YMYassine Mekdad
Yassine Mekdad
GBGiuseppe Bernieri
Giuseppe Bernieri
MCMauro Conti
Mauro Conti
AFAbdeslam El Fergougui
Abdeslam El Fergougui

12 February 2025
Verified
Conference Paper
Open access
FBFernando Brito
Fernando Brito
YMYassine Mekdad
Yassine Mekdad
MRMonique Ross
Monique Ross
MFMark A. Finlayson
Mark A. Finlayson
SUSelcuk Uluagac
Selcuk Uluagac
Artificial Intelligence (AI) has become a fundamental tool for cybersecurity researchers and practitioners. It is frequently used to address major security problems such as supply chain attacks, ransomware threats, and social engineering. In this context, integrating AI into cybersecurity workflows requires incorporating AI-driven approaches into the educational training of the cybersecurity workforce. This paradigm shift in academic settings will introduce the necessary skills for cybersecurity professionals to operate modern AI-based systems. Yet, the current cybersecurity curriculum still suffers from the absence of AI resources, particularly the detailed understanding of the appropriate AI mechanisms. Such absence leaves skill gaps for future professionals and practitioners in the industry. To address this, we designed an academic lecture module on AI covering both theory and practice. Then, we taught the module across six cybersecurity courses in our institution. To assess the effectiveness of integrating AI materials into cybersecurity education, we collected data by presenting two surveys before and after the lecture (concluding 81 participants per survey). Specifically, we utilized widely accepted models for unbiased analysis of our data. Our experimental results show positive AI knowledge improvement by 30% of the participants, demonstrating the beneficial impact of the lecture. Then, we observed a high similarity score between the survey responses and the lecture content, reaching 84%. Moreover, our sentiment analysis results reflect positive feedback from the participants with a positive score of 0.50. Overall, our study serves as a reference for instructional designers for developing educational curricula aiming to integrate AI into cybersecurity education.