TY  - RPRT
TI  - Non-convergence of stochastic gradient descent in the training of deep neural networks
AU  - Patrick Cheridito
AU  - Arnulf Jentzen
AU  - Florian Rossmannek
PY  - 2021
DO  - 10.1016/j.jco.2020.101540
UR  - https://arxiv.org/abs/2006.07075
ID  - 2006.07075
ER  -