Further investigation into multilingual training and adaptation of stacked bottle-neck neural network structure.
Frantisek GrézlEkaterina EgorovaMartin KarafiátPublished in: SLT (2014)
Keyphrases
- neural network structure
- bottle neck
- training process
- neural network
- training algorithm
- fuzzy neural network
- hidden layer
- recurrent neural networks
- multi layer perceptron
- digital libraries
- machine learning
- regression model
- connection weights
- back propagation
- artificial neural networks
- learning rate
- training set
- training data
- data flow
- social networks
- genetic algorithm