CUDA: Convolution-based Unlearnable Datasets. (arXiv:2303.04278v1 [cs.LG])

Large-scale training of modern deep learning models heavily relies on
publicly available data on the web. This potentially unauthorized usage of
online data leads to concerns regarding data privacy. Recent works aim to make
unlearnable data for deep learning models by adding small, specially designed
noises to tackle this issue. However, these methods are vulnerable to
adversarial training (AT) and/or are computationally heavy. In this work, we
propose a novel, model-free, Convolution-based Unlearnable DAtaset (CUDA)
generation technique. CUDA is generated …

adversarial aim data data privacy datasets deep learning free issue large novel privacy scale the web training vulnerable web work