Few-shot Backdoor Attacks via Neural Tangent Kernels
This work addresses security vulnerabilities in deep neural networks for applications like image classification, though it is incremental as it builds on existing backdoor attack methods with a new optimization approach.
The authors tackled the problem of backdoor attacks in deep learning by formulating it as a bilevel optimization problem and using neural tangent kernels to design strong poison examples, achieving a 90% attack success rate with ten times fewer poison examples than baselines on datasets like CIFAR-10 and ImageNet.
In a backdoor attack, an attacker injects corrupted examples into the training set. The goal of the attacker is to cause the final trained model to predict the attacker's desired target label when a predefined trigger is added to test inputs. Central to these attacks is the trade-off between the success rate of the attack and the number of corrupted training examples injected. We pose this attack as a novel bilevel optimization problem: construct strong poison examples that maximize the attack success rate of the trained model. We use neural tangent kernels to approximate the training dynamics of the model being attacked and automatically learn strong poison examples. We experiment on subclasses of CIFAR-10 and ImageNet with WideResNet-34 and ConvNeXt architectures on periodic and patch trigger attacks and show that NTBA-designed poisoned examples achieve, for example, an attack success rate of 90% with ten times smaller number of poison examples injected compared to the baseline. We provided an interpretation of the NTBA-designed attacks using the analysis of kernel linear regression. We further demonstrate a vulnerability in overparametrized deep neural networks, which is revealed by the shape of the neural tangent kernel.