On Learnability via Gradient Method
for Two-Layer ReLU Neural Networks in Teacher-Student Setting
Shunta Akiyama 1 Taiji Suzuki 1 2
Abstract for the generalization aspect. In this study, we tackle these
two problems in a teacher-student problem with the ReLU
Deep learning empirically achieves high perfor- activation under an over-parameterized setting. In this set-
mance in many applications, bu ...


雷达卡




京公网安备 11010802022788号







