Provable Identifiability of Two-Layer ReLU Neural Networks via LASSO Regularization.
Gen LiGanghua WangJie DingPublished in: IEEE Trans. Inf. Theory (2023)
Keyphrases
- neural network
- multi layer
- regularization parameter
- elastic net
- feature selection
- pattern recognition
- group lasso
- single layer
- artificial neural networks
- cross validation
- variable selection
- fuzzy logic
- back propagation
- regularization method
- image restoration
- neural network model
- multilayer perceptron
- neural nets
- model selection
- regularization term
- genetic algorithm
- recurrent neural networks
- linear regression
- causal models
- machine learning
- support vector
- application layer
- least squares
- fault diagnosis
- sample size
- loss function