TY  - RPRT
TI  - A proof of convergence for gradient descent in the training of artificial neural networks for constant target functions
AU  - Patrick Cheridito
AU  - Arnulf Jentzen
AU  - Adrian Riekert
AU  - Florian Rossmannek
PY  - 2021
DO  - 10.1016/j.jco.2022.101646
UR  - https://arxiv.org/abs/2102.09924
ID  - 2102.09924
ER  -