发帖
雷达卡
Train simultaneously, generalize better:
Stability of gradient-based minimax learners
Farzan Farnia 1 Asuman Ozdaglar 1
Abstract 2014) and adversarial training (Madry et al., 2017) have
achieved great success over a wide array of learning tasks.
The success of minimax learning problems of
In these approaches, the learning problem is modeled as a
generative adv ...
京公网安备 11010802022788号
