Keyphrases
CIFAR-10
50%
Convolutional
50%
Deep Learning
50%
Depth Limit
50%
Dynamic Limit
100%
Dynamical Mean-field Theory
50%
Feature Learning
50%
Finite Size
50%
Hyperparameter Tuning
50%
Hyperparameters
100%
ImageNet
50%
Infinite Depth
50%
Large Width
50%
Learning Dynamics
50%
Model Size
50%
Network Dynamics
100%
Network Scale
100%
Network Transfer
50%
Neural Network Learning
50%
Optimal Hyperparameters
100%
Residual Architecture
50%
Residual Network
100%
Scaling Limit
100%
Small Networks
50%
Tuning Methods
50%
U-Net
100%
Vision Transformer
50%
Well-defined
50%
Mathematics
Deep Learning Method
50%
Network Dynamic
100%
Neural Network
50%
Representation Learning
50%
Residual Network
100%
Residuals
100%
Scaling Limit
100%
Chemical Engineering
Deep Learning Method
100%
Neural Network
100%
Biochemistry, Genetics and Molecular Biology