Rect-ViT: Rectified attention via feature attribution can improve the adversarial robustness of Vision Transformers

作者：Xu Kang, Bin Song · 发表于：Neural Networks · 年份：2025 · DOI：10.1016/j.neunet.2025.107666 · 被引用次数：3 · 研究领域：Adversarial Robustness in Machine Learning、Advanced Neural Network Applications、Domain Adaptation and Few-Shot Learning