On the difficulty of training recurrent neural networks.
Razvan PascanuTomás MikolovYoshua BengioPublished in: ICML (3) (2013)
Keyphrases
- recurrent neural networks
- recurrent networks
- echo state networks
- feedforward neural networks
- neural network
- feed forward
- artificial neural networks
- reservoir computing
- hidden layer
- complex valued
- feed forward neural networks
- neural model
- cascade correlation
- hebbian learning
- long short term memory
- chaotic time series
- training examples
- training samples
- machine learning
- training algorithm
- neural models
- supervised learning
- nonlinear dynamic systems
- training set
- decision making