Convergence to good non-optimal critical points in the training of neural networks: Gradient descent optimization with one random initialization overcomes all bad non-global local minima with high probability
Ibragimov, S.; Jentzen, A.; Riekert, A.
Research article in digital collection | Preprint | Peer reviewedDetails about the publication
Name of the repository: arXiv.org
Article number: 2212.13111
Status: submitted / under review
Release year: 2022
Link to the full text: https://arxiv.org/abs/2212.13111
Authors from the University of Münster
| Ibragimov, Shokhrukh | Professorship for applied mathematics (Prof. Jentzen) |
| Jentzen, Arnulf | Institute for Analysis and Numerics |
| Riekert, Adrian | Mathematical Institute |