Robustifying Token Attention for Vision Transformers | IEEE Conference Publication | IEEE Xplore