Noisy labels are ubiquitous in real-world datasets, which poses a challenge for robustly
training deep neural networks (DNNs) as DNNs usually have the high capacity to memorize …

Deep Neural Networks (DNNs) have recently been shown to be vulnerable against
adversarial examples, which are carefully crafted instances that can mislead DNNs to make …

Datasets with significant proportions of noisy (incorrect) class labels present challenges for
training accurate Deep Neural Networks (DNNs). We propose a new perspective for …

Deep neural networks progressively transform their inputs across multiple processing layers.
What are the geometrical properties of the representations learned by these networks? Here …

Obtaining deep networks that are robust against adversarial examples and generalize well
is an open problem. A recent hypothesis even states that both robust and accurate models …

Adversarial training yields robust models against a specific threat model, eg, $ L_\infty $
adversarial examples. Typically robustness does not generalize to previously unseen threat …

Data poisoning attacks and backdoor attacks aim to corrupt a machine learning classifier via
modifying, adding, and/or removing some carefully selected training examples, such that the …

Adversarial training (AT) has become the de-facto standard to obtain models robust against
adversarial examples. However, AT exhibits severe robust overfitting: cross-entropy loss on …

Motivated by safety-critical applications, test-time attacks on classifiers via adversarial
examples has recently received a great deal of attention. However, there is a general lack of …

Geometric descriptions of deep neural networks (DNNs) have the potential to uncover core
representational principles of computational models in neuroscience. Here we examined the …