TY  - RPRT
TI  - Fast generalization error bound of deep learning without scale invariance of activation functions
AU  - Yoshikazu Terada
AU  - Ryoma Hirose
PY  - 2019
UR  - https://arxiv.org/abs/1907.10900
ID  - 1907.10900
ER  -