Integration of Unsupervised and Supervised Criteria for Deep Neural Networks Training.
Francisco Zamora-MartínezJavier Muñoz-AlmarazJuan PardoPublished in: ICANN (2) (2016)
Keyphrases
- supervised learning
- neural network
- supervised training
- deep architectures
- unsupervised learning
- supervised methods
- semi supervised
- training process
- training algorithm
- neural nets
- restricted boltzmann machine
- deep learning
- unsupervised methods
- pattern recognition
- feedforward neural networks
- back propagation
- feed forward neural networks
- supervised classification
- deep belief networks
- training phase
- training set
- training examples
- weakly supervised
- backpropagation algorithm
- fault diagnosis
- fuzzy logic
- error back propagation
- activation function
- multi layer perceptron
- unsupervised feature selection
- discriminant projection
- multi layer
- neural network model
- training data
- artificial neural networks
- training samples
- evaluation criteria
- competitive learning
- hidden layer
- machine learning
- learning algorithm
- neural network training
- sufficient conditions
- self organizing maps
- multilayer perceptron
- labeled training data
- labeled data
- semi supervised learning
- global exponential stability
- expectation maximization
- test set
- feature space
- support vector
- data integration
- feed forward
- genetic algorithm