Cargando…
A mean field view of the landscape of two-layer neural networks
Multilayer neural networks are among the most powerful models in machine learning, yet the fundamental reasons for this success defy mathematical understanding. Learning a neural network requires optimizing a nonconvex high-dimensional objective (risk function), a problem that is usually attacked us...
Autores principales: | Mei, Song, Montanari, Andrea, Nguyen, Phan-Minh |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
National Academy of Sciences
2018
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6099898/ https://www.ncbi.nlm.nih.gov/pubmed/30054315 http://dx.doi.org/10.1073/pnas.1806579115 |
Ejemplares similares
-
Shaping the learning landscape in neural networks around wide flat minima
por: Baldassi, Carlo, et al.
Publicado: (2020) -
High-resolution limited-angle phase tomography of dense layered objects using deep neural networks
por: Goy, Alexandre, et al.
Publicado: (2019) -
The immunopeptidomic landscape of ovarian carcinomas
por: Schuster, Heiko, et al.
Publicado: (2017) -
Resonance with subthreshold oscillatory drive organizes activity and optimizes learning in neural networks
por: Roach, James P., et al.
Publicado: (2018) -
Complex modifier landscape underlying genetic background effects
por: Hou, Jing, et al.
Publicado: (2019)