One Neuron to Fool Them All
Despite vast research in adversarial examples, the root causes of model susceptibility are not well understood. Instead of looking at attack-specific robustness, we propose a notion that evaluates the sensitivity of individual neurons in terms of how robust the model's output is to direct perturbations of that neuron's output. Analyzing models from this perspective reveals distinctive characteristics of standard as well as adversarially-trained robust models, and leads to several curious results. In our experiments on CIFAR-10 and ImageNet, we find that attacks using a loss function that targets just a single sensitive neuron find adversarial examples nearly as effectively as ones that target the full model. We analyze the properties of these sensitive neurons to propose a regularization term that can help a model achieve robustness to a variety of different perturbation constraints while maintaining accuracy on natural data distributions. Code for all our experiments is available at https://github.com/iamgroot42/sauron .
Code (1)
Tasks
AllSimilar Papers 제목 키워드 기반
NAT: Learning to Attack Neurons for Enhanced Adversarial Transferability
The generation of transferable adversarial perturbations typically involves training a generator to maximize embedding separation between clean and adversarial images at a single mid-layer of a source model. In this work…
FooBaR: Fault Fooling Backdoor Attack on Neural Network Training
Neural network implementations are known to be vulnerable to physical attack vectors such as fault injection attacks. As of now, these attacks were only utilized during the inference phase with the intention to cause a m…
Backdoor Attackimage-classificationImage ClassificationVisual Analytics of Neuron Vulnerability to Adversarial Attacks on Convolutional Neural Networks
Adversarial attacks on a convolutional neural network (CNN) -- injecting human-imperceptible perturbations into an input image -- could fool a high-performance CNN into making incorrect predictions. The success of advers…
Autonomous DrivingMedical DiagnosisVisual ReasoningFinite Gaussian Neurons: Defending against adversarial attacks by making neural networks say "I don't know"
Since 2014, artificial neural networks have been known to be vulnerable to adversarial attacks, which can fool the network into producing wrong or nonsensical outputs by making humanly imperceptible alterations to inputs…
Fight Perturbations with Perturbations: Defending Adversarial Attacks via Neuron Influence
The vulnerabilities of deep learning models towards adversarial attacks have attracted increasing attention, especially when models are deployed in security-critical domains. Numerous defense methods, including reactive …