Neural tangent kernels, transportation mappings, and universal approximation.
Ziwei JiMatus TelgarskyRuicheng XianPublished in: CoRR (2019)
Keyphrases
- neural network
- network architecture
- approximation error
- error bounds
- approximation algorithms
- transportation systems
- gram matrix
- kernel methods
- kernel function
- relative error
- support vector
- closed form
- linear combination
- feature space
- bio inspired
- neural model
- transportation problem
- positive definite
- approximation methods
- gaussian kernels
- multiscale
- approximation schemes
- decision making