Sobolev Training with Approximated Derivatives for Black-Box Function Regression with Neural Networks.
Matthias KisselKlaus DiepoldPublished in: ECML/PKDD (2) (2019)
Keyphrases
- black box
- neural network
- rule extraction
- hybrid systems
- training process
- black boxes
- nonlinear functions
- radial basis function network
- training algorithm
- white box
- feedforward neural networks
- integration testing
- linear combination
- multi layer perceptron
- backpropagation algorithm
- fuzzy logic
- regression function
- feed forward neural networks
- state transition
- activation function
- support vector
- test set
- training set
- test cases
- case study
- model selection
- machine learning
- artificial neural networks
- databases
- training data
- regression model
- decision trees
- training samples
- metadata
- artificial intelligence
- supervised learning