Check Your Other Door! Creating Backdoor Attacks in the Frequency Domain. (arXiv:2109.05507v3 [cs.CR] UPDATED)

Deep Neural Networks (DNNs) are ubiquitous and span a variety of applications
ranging from image classification to real-time object detection. As DNN models
become more sophisticated, the computational cost of training these models
becomes a burden. For this reason, outsourcing the training process has been
the go-to option for many DNN users. Unfortunately, this comes at the cost of
vulnerability to backdoor attacks. These attacks aim to establish hidden
backdoors in the DNN so that it performs well on clean …

aim applications attacks backdoor backdoor attacks check classification computational cost detection domain door networks neural networks object outsourcing process span training vulnerability