Robust Overfitting may be mitigated by properly learned smoothening.

Tianlong Chen, Zhenyu Zhang, Sijia Liu, Shiyu Chang, Zhangyang Wang

Browse the full ICLR paper archive.