ResNet with one-neuron hidden layers is a Universal Approximator.
Hongzhou LinStefanie JegelkaPublished in: CoRR (2018)
Keyphrases
- hidden layer
- activation function
- output layer
- neural network
- feed forward
- back propagation
- learning rate
- artificial neural networks
- input output
- feed forward neural networks
- feedforward neural networks
- radial basis function
- rbf neural network
- multilayer perceptron
- recurrent neural networks
- neural nets
- training algorithm
- latent variables
- backpropagation algorithm
- hidden nodes
- number of hidden layers
- multi layer perceptron
- learning algorithm
- basis functions
- hidden units
- feature extraction
- artificial intelligence
- cluster analysis
- bayesian networks
- machine learning