Global optimality of softmax policy gradient with single hidden layer neural networks in the mean-field regime.
Andrea AgazziJianfeng LuPublished in: ICLR (2021)
Keyphrases
- activation function
- neural network
- global optimality
- global optimization
- hidden layer
- feed forward
- back propagation
- artificial neural networks
- multilayer perceptron
- optimal solution
- markov random field
- learning rate
- radial basis function
- hidden nodes
- basis functions
- em algorithm
- objective function
- fuzzy logic
- global minimum
- fuzzy neural network
- markov networks
- recurrent neural networks
- input space
- rbf neural network
- genetic algorithm
- bayesian networks
- probabilistic model
- optimization methods
- fixed point