Mitigating the Impact of Adversarial Attacks in Very Deep Networks
This research provides an incremental defense mechanism against data poisoning attacks for deep neural networks, which is relevant for practitioners concerned with model security and robustness.
This paper addresses the vulnerability of deep neural networks to data poisoning attacks, which degrade model accuracy and convergence. The authors propose an attack-agnostic defense method that integrates a Defensive Feature Layer (DFL) and a Polarized Contrastive Loss (PCL) to neutralize illegitimate perturbation samples and improve discrimination, showing excellent performance against recent peer techniques on CIFAR-10 and MNIST datasets.
Deep Neural Network (DNN) models have vulnerabilities related to security concerns, with attackers usually employing complex hacking techniques to expose their structures. Data poisoning-enabled perturbation attacks are complex adversarial ones that inject false data into models. They negatively impact the learning process, with no benefit to deeper networks, as they degrade a model's accuracy and convergence rates. In this paper, we propose an attack-agnostic-based defense method for mitigating their influence. In it, a Defensive Feature Layer (DFL) is integrated with a well-known DNN architecture which assists in neutralizing the effects of illegitimate perturbation samples in the feature space. To boost the robustness and trustworthiness of this method for correctly classifying attacked input samples, we regularize the hidden space of a trained model with a discriminative loss function called Polarized Contrastive Loss (PCL). It improves discrimination among samples in different classes and maintains the resemblance of those in the same class. Also, we integrate a DFL and PCL in a compact model for defending against data poisoning attacks. This method is trained and tested using the CIFAR-10 and MNIST datasets with data poisoning-enabled perturbation attacks, with the experimental results revealing its excellent performance compared with those of recent peer techniques.