Backpropagation through the Void: Optimizing control variates for black-box gradient estimation.
Will GrathwohlDami ChoiYuhuai WuGeoffrey RoederDavid DuvenaudPublished in: ICLR (Poster) (2018)
Keyphrases
- black box
- back propagation
- gradient estimation
- artificial neural networks
- neural network
- black boxes
- feed forward
- training algorithm
- neural nets
- backpropagation neural network
- multilayer perceptron
- white box
- hidden layer
- control system
- backpropagation neural networks
- feed forward neural networks
- feedforward neural networks
- bp neural network
- bp algorithm
- backpropagation algorithm
- learning algorithm
- neural network ensemble
- white box testing
- fuzzy logic
- rule extraction
- real time
- test cases
- variance reduction
- integration testing
- multilayer perceptron neural network
- multilayer neural network
- activation function
- reinforcement learning
- optimal control
- support vector machine