Despite recent success, state-of-the-art learning-based models remain hi...
Privacy-preserving machine learning aims to train models on private data...
In safety-critical classification tasks, conformal prediction allows to
...
For safety, AI systems in health undergo thorough evaluations before
dep...
Despite the success of vision transformers (ViTs), they still suffer fro...
Modern deep learning architecture utilize batch normalization (BN) to
st...
Deep neural networks have achieved great success in many computer vision...
Modern deep learning based classifiers show very high accuracy on test d...
We study the adversarial robustness of information bottleneck models for...
Deep neural network (DNN) accelerators received considerable attention i...
Adversarial training (AT) has become the de-facto standard to obtain mod...
Deep neural network (DNN) accelerators received considerable attention i...
Deep neural networks have been shown to be susceptible to adversarial
ex...
Adversarial training is the standard to train models robust against
adve...
Obtaining deep networks that are robust against adversarial examples and...
We address the problem of 3D shape completion from sparse and noisy poin...
Superpixels group perceptually similar pixels to create visually meaning...