Efficient Training of Context-Dependent Neural Nets with Conjugate Gradient Algorithms.
Piotr CiskowskiPublished in: DepCoS-RELCOMEX (2014)
Keyphrases
- active learning
- learning algorithm
- conjugate gradient
- context dependent
- neural nets
- training algorithm
- back propagation
- neural network
- training samples
- convergence rate
- maximum likelihood estimation
- data dependent
- special case
- low level
- expectation maximization
- feature selection
- artificial neural networks
- pairwise
- training process
- expert systems
- reproducing kernel hilbert space
- decision trees