Tight Bounds on the Smallest Eigenvalue of the Neural Tangent Kernel for Deep ReLU Networks.
Quynh NguyenMarco MondelliGuido F. MontúfarPublished in: ICML (2021)
Keyphrases
- tight bounds
- upper bound
- neural network
- social networks
- network architecture
- bio inspired
- least squares
- kernel methods
- network model
- computer networks
- covariance matrix
- network structure
- data sets
- kernel function
- dimensionality reduction
- feature space
- network design
- component analysis
- kernel pca
- kernel regression
- learning algorithm
- recurrent networks