Abstract
One of the major concerns for neural network training is that the nonconvexity of the associated loss functions may cause a bad landscape. The recent success of neural networks suggests that their loss landscape is not too bad, but what specific results do we know about the landscape? In this article, we review recent findings and results on the global landscape of neural networks.
Original language | English (US) |
---|---|
Article number | 9194023 |
Pages (from-to) | 95-108 |
Number of pages | 14 |
Journal | IEEE Signal Processing Magazine |
Volume | 37 |
Issue number | 5 |
DOIs | |
State | Published - Sep 2020 |
ASJC Scopus subject areas
- Signal Processing
- Electrical and Electronic Engineering
- Applied Mathematics