Cargando…

Studying the Evolution of Neural Activation Patterns During Training of Feed-Forward ReLU Networks

The ability of deep neural networks to form powerful emergent representations of complex statistical patterns in data is as remarkable as imperfectly understood. For deep ReLU networks, these are encoded in the mixed discrete–continuous structure of linear weight matrices and non-linear binary activ...

Descripción completa

Detalles Bibliográficos
Autores principales: Hartmann, David, Franzen, Daniel, Brodehl, Sebastian
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Frontiers Media S.A. 2021
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8733739/
https://www.ncbi.nlm.nih.gov/pubmed/35005614
http://dx.doi.org/10.3389/frai.2021.642374