We also apply PGD-1 and PGD-2 w/wo FN to attack a standard WRN-34-10 model, and

Neural Information Processing Systems 

We thank all the reviewers for their valuable comments. Below, we address the detailed comments of each reviewer. The margin range m is chosen to be around cos 30 0.15. Then other samples that are not well-learned will dynamically contribute more. Figure 1, Sec. 3.5, and other comments: Thank you for the suggestions.