TY  - RPRT
TI  - A single gradient step finds adversarial examples on random two-layers neural networks
AU  - Sébastien Bubeck
AU  - Yeshwanth Cherapanamjeri
AU  - Gauthier Gidel
AU  - Rémi Tachet des Combes
PY  - 2021
UR  - https://arxiv.org/abs/2104.03863
ID  - 2104.03863
ER  -