A single-layer RNN can approximate stacked and bidirectional RNNs, and topologies in between.
Javier S. TurekShailee JainMihai CapotaAlexander G. HuthTheodore L. WillkePublished in: CoRR (2019)
Keyphrases
- recurrent neural networks
- single layer
- hidden layer
- neural network
- feed forward
- neural nets
- feed forward neural networks
- long short term memory
- activation function
- multi layer
- feedforward neural networks
- multiple layers
- artificial neural networks
- training algorithm
- inter layer
- fuzzy neural network
- neuro fuzzy
- scalable video coding
- back propagation
- machine learning
- hopfield neural network
- hidden units
- markov random field