Elaf Abdullah Makkawi
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Elaf Abdullah Makkawi has a total of 5 co-authors. Here are the top co-authors by number of shared publications:
- Ali Saeed Almuflih
Department of Industrial Engineering, College of Engineering, King Khalid University, Abha, Saudi Arabia
1 shared publication - Dhairya Vyas
The Maharaja Sayajirao University of Baroda
1 shared publication - Karishma Mohamed Rafik Qureshi
Department of Mechanical Engineering, Parul University, Waghodia 391760, India
1 shared publication - Mohamed Rafik Noor Mohamed Qureshi
Department of Industrial Engineering, College of Engineering, King Khalid University, Abha, Saudi Arabia
1 shared publication - Viral Vinod Kapadia
Computer Science and Engineering Department, The Maharaja Sayajirao University of Baroda, Vadodara 390002, India
1 shared publication
Latest Publications
- Open Access
- 20 May 2022
- Published by MDPI AG in Applied Sciences
- Vol. 12 (10) , 5161
Novel Exploit Feature-Map-Based Detection of Adversarial Attacks
In machine learning (ML), adversarial attack (targeted or untargeted) in the presence of noise disturbs the model prediction. This research suggests that adversarial perturbations on pictures lead to noise in the features constructed by any networks. As a result, adversarial assaults against image categorization systems may present obstacles and possibilities for studying convolutional neural networks (CNNs). According to this research, adversarial perturbations on pictures cause noise in the features created by neural networks. Motivated by adversarial perturbation on image pixel attacks observation, we developed a novel exploit feature map that describes adversarial attacks by performing individual object feature-map visual description. Specifically, a novel detection algorithm calculates each object’s class activation map weight and makes a combined activation map. When checked with different networks like VGGNet19 and ResNet50, in both white-box and black-box attack situations, the unique exploit feature-map significantly improves the state-of-the-art in adversarial resilience. Further, it will clearly exploit attacks on ImageNet under various algorithms like Fast Gradient Sign Method (FGSM), DeepFool, Projected Gradient Descent (PGD), and Backward Pass Differentiable Approximation (BPDA).
- 20 May 2022
- Published by MDPI AG in Applied Sciences
- Vol. 12 (10) , 5161
Open AccessNovel Exploit Feature-Map-Based Detection of Adversarial Attacks
In machine learning (ML), adversarial attack (targeted or untargeted) in the presence of noise disturbs the model prediction. This research suggests that adversarial perturbations on pictures lead to noise in the features constructed by any networks. As a result, adversarial assaults against image categorization systems may present obstacles and possibilities for studying convolutional neural networks (CNNs). According to this research, adversarial perturbations on pictures cause noise in the features created by neural networks. Motivated by adversarial perturbation on image pixel attacks observation, we developed a novel exploit feature map that describes adversarial attacks by performing individual object feature-map visual description. Specifically, a novel detection algorithm calculates each object’s class activation map weight and makes a combined activation map. When checked with different networks like VGGNet19 and ResNet50, in both white-box and black-box attack situations, the unique exploit feature-map significantly improves the state-of-the-art in adversarial resilience. Further, it will clearly exploit attacks on ImageNet under various algorithms like Fast Gradient Sign Method (FGSM), DeepFool, Projected Gradient Descent (PGD), and Backward Pass Differentiable Approximation (BPDA).