Forcing Interpretability for Deep Neural Networks through Rule-Based Regularization.
Nadia BurkartMarco F. HuberPhillip FallerPublished in: ICMLA (2019)
Keyphrases
- neural network
- rule base
- pattern recognition
- fuzzy logic
- expert systems
- fuzzy systems
- rule based systems
- back propagation
- neural network model
- neural nets
- prediction accuracy
- data driven
- artificial neural networks
- regularization parameter
- multilayer perceptron
- parameter selection
- prior information
- data sets
- feed forward
- recurrent neural networks
- training data
- hidden layer
- training algorithm
- image processing
- regularization framework