Sampling-Based Gradient Regularization for Capturing Long-Term Dependencies in Recurrent Neural Networks.
Artem N. ChernodubDimitri NowickiPublished in: ICONIP (2) (2016)
Keyphrases
- recurrent neural networks
- long term
- short term
- feed forward
- complex valued
- neural network
- artificial neural networks
- recurrent networks
- neural model
- motion planning
- echo state networks
- reservoir computing
- cascade correlation
- feedforward neural networks
- nonlinear dynamic systems
- long short term memory
- regularization term
- monte carlo
- prior information
- image gradient
- data mining
- piecewise constant
- chaotic time series
- generative model
- edge detection
- probabilistic model