TY  - RPRT
TI  - Quantifying the Preferential Direction of the Model Gradient in Adversarial Training With Projected Gradient Descent
AU  - Ricardo Bigolin Lanfredi
AU  - Joyce D. Schroeder
AU  - Tolga Tasdizen
PY  - 2023
DO  - 10.1016/j.patcog.2023.109430
UR  - https://arxiv.org/abs/2009.04709
ID  - 2009.04709
ER  -