Results 231 to 240 of about 217,331 (267)

Boosting adversarial robustness via self-paced adversarial training

Neural Networks, 2023
Adversarial training is considered one of the most effective methods to improve the adversarial robustness of deep neural networks. Despite the success, it still suffers from unsatisfactory performance and overfitting. Considering the intrinsic mechanism of adversarial training, recent studies adopt the idea of curriculum learning to alleviate ...
Lirong He   +5 more
openaire   +2 more sources

Adversarial Training With Anti-Adversaries

IEEE Transactions on Pattern Analysis and Machine Intelligence
Adversarial training is effective in improving the robustness of deep neural networks. However, existing studies still exhibit significant drawbacks in terms of the robustness, generalization, and fairness of models. In this study, we validate the importance of different perturbation directions (i.e., adversarial and anti-adversarial) and bounds from ...
Xiaoling Zhou, Ou Wu, Nan Yang
openaire   +2 more sources

Adversarial training with Lookahead

2022
Deep Learning wird in immer mehr sicherheitsrelevanten Bereichen wie zum Beispiel für autonomes Fahren oder in der automatischen Gesichtserkennung erfolgreich eingesetzt. Daher rückt der Sicherheitsaspekt von Deep Learning Algorithmen zunehmend in den Fokus der Forschung.
openaire   +1 more source

Adversarial training

2023
Pin-Yu Chen, Cho-Jui Hsieh
openaire   +1 more source

Variational Adversarial Defense: A Bayes Perspective for Adversarial Training

IEEE Transactions on Pattern Analysis and Machine Intelligence
Various methods have been proposed to defend against adversarial attacks. However, there is a lack of enough theoretical guarantee of the performance, thus leading to two problems: First, deficiency of necessary adversarial training samples might attenuate the normal gradient's back-propagation, which leads to overfitting and gradient masking ...
Chenglong Zhao   +5 more
openaire   +2 more sources

Home - About - Disclaimer - Privacy