Scholarlink

<p>Deep neural networks (DNNs) have achieved remarkable success in computer vision tasks such as image classification, segmentation, and object detection. However, they are vulnerable to adversarial attacks, which can cause incorrect predictions with small perturbations in input images. Addressing this issue is crucial for deploying robust deep-learning systems. This paper presents a novel approach that utilizes contrastive learning for adversarial defense, a previously unexplored area. Our method leverages the contrastive loss function to enhance the robustness of classification models by training them with both clean and adversarially perturbed images. The framework utilizes a ResNet-based encoder and a two-layer projection head, and it leverages multiple adversarial variants generated using FGSM, PGD, and CW attacks to form positive pairs for contrastive training. By optimizing the model’s parameters alongside the perturbations, our approach enables the network to learn robust representations that are less susceptible to adversarial attacks. Experimental results show significant improvements in the model’s robustness against various types of adversarial perturbations. Evaluated on the CIFAR-10 dataset, C-LEAD demonstrates improved robustness, achieving 68.67% accuracy under PGD and 72.43% under FGSM, surpassing conventional adversarial training methods. This suggests that contrastive loss helps extract more informative and resilient features, contributing to the field of adversarial robustness in deep learning. The code is publicly made available on github in the following link: <a href="https://github.com/suklav/C_Lead">https://github.com/suklav/C_Lead.</a></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

C-LEAD: Contrastive Learning for Enhanced Adversarial Defense

  • 期刊论文

  • 发表日期：

    2026年5月29日

查看全文
  • Suklav Ghosh,
  • Sonal Kumar,
  • Arijit Sur

摘要

Deep neural networks (DNNs) have achieved remarkable success in computer vision tasks such as image classification, segmentation, and object detection. However, they are vulnerable to adversarial attacks, which can cause incorrect predictions with small perturbations in input images. Addressing this issue is crucial for deploying robust deep-learning systems. This paper presents a novel approach that utilizes contrastive learning for adversarial defense, a previously unexplored area. Our method leverages the contrastive loss function to enhance the robustness of classification models by training them with both clean and adversarially perturbed images. The framework utilizes a ResNet-based encoder and a two-layer projection head, and it leverages multiple adversarial variants generated using FGSM, PGD, and CW attacks to form positive pairs for contrastive training. By optimizing the model’s parameters alongside the perturbations, our approach enables the network to learn robust representations that are less susceptible to adversarial attacks. Experimental results show significant improvements in the model’s robustness against various types of adversarial perturbations. Evaluated on the CIFAR-10 dataset, C-LEAD demonstrates improved robustness, achieving 68.67% accuracy under PGD and 72.43% under FGSM, surpassing conventional adversarial training methods. This suggests that contrastive loss helps extract more informative and resilient features, contributing to the field of adversarial robustness in deep learning. The code is publicly made available on github in the following link: https://github.com/suklav/C_Lead.