Input-Cell Attention Reduces Vanishing Saliency of Recurrent Neural Networks.
Aya Abdelsalam IsmailMohamed K. GunadyLuiz PessoaHéctor Corrada BravoSoheil FeiziPublished in: CoRR (2019)
Keyphrases
- recurrent neural networks
- visual attention
- saliency map
- feed forward
- neural network
- recurrent networks
- long short term memory
- echo state networks
- feedforward neural networks
- reservoir computing
- visual saliency
- hidden layer
- complex valued
- attention mechanism
- input data
- nonlinear dynamic systems
- artificial neural networks
- cascade correlation
- data mining
- neural model
- salient regions
- hebbian learning
- focus of attention
- vision system
- information content
- eye tracking
- long term
- expert systems