Gradient-Based Adversarial and Out-of-Distribution Detection. (arXiv:2206.08255v1 [cs.LG])

Web: http://arxiv.org/abs/2206.08255

We propose to utilize gradients for detecting adversarial and
out-of-distribution samples. We introduce confounding labels -- labels that
differ from normal labels seen during training -- in gradient generation to
probe the effective expressivity of neural networks. Gradients depict the
amount of change required for a model to properly represent given inputs,
providing insight into the representational power of the model established by
network architectural properties as well as training data. By introducing a
label of different design, we remove …

arxiv detection distribution gradient lg