Cargando…
Shallow Univariate ReLU Networks as Splines: Initialization, Loss Surface, Hessian, and Gradient Flow Dynamics
Understanding the learning dynamics and inductive bias of neural networks (NNs) is hindered by the opacity of the relationship between NN parameters and the function represented. Partially, this is due to symmetries inherent within the NN parameterization, allowing multiple different parameter setti...
Autores principales: | Sahs, Justin, Pyle, Ryan, Damaraju, Aneel, Caro, Josue Ortega, Tavaslioglu, Onur, Lu, Andy, Anselmi, Fabio, Patel, Ankit B. |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
Frontiers Media S.A.
2022
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9131019/ https://www.ncbi.nlm.nih.gov/pubmed/35647529 http://dx.doi.org/10.3389/frai.2022.889981 |
Ejemplares similares
-
Studying the Evolution of Neural Activation Patterns During Training of Feed-Forward ReLU Networks
por: Hartmann, David, et al.
Publicado: (2021) -
Understanding Robustness and Generalization of Artificial Neural Networks Through Fourier Masks
por: Karantzas, Nikos, et al.
Publicado: (2022) -
An Empirical Investigation Into Deep and Shallow Rule Learning
por: Beck, Florian, et al.
Publicado: (2021) -
Integrating geometries of ReLU feedforward neural networks
por: Liu, Yajing, et al.
Publicado: (2023) -
Training a Two-Layer ReLU Network Analytically
por: Barbu, Adrian
Publicado: (2023)