Generator functions and gradient descent learning

10. Application: controlling water reservoirs
10.1 Implementation results
10.2 Rule extraction
11. Application of the statistical approach
Discussion 12.

References

1. Introduction
2. Function approximation models and RBF neural networks 3. Reformulating radial basis neural networks
4. Admissible generator functions
4.1 Linear generator functions
4.2 Exponential generator functions
5. Selecting generatorfunctions
5.1 The blind spot
5.2 Criteria for selecting generator functions
5.3 Evaluation of linear and exponential generator functions 5.3.1 Linear generator functions
5.3.2 Exponential generator functions
6. Learning algorithms based on gradient descent
6.1 Batch learning algorithms
6.2 Sequential learning algorithms
6.3 Initialization of supervised learning
7. Generator functions and gradient descent learning
8. Handwritten digit recognition
8.1 The NIST databases
8.2 Data preprocessing
8.3 Classification tools for NIST digits
8.4 Role of the prototypes in gradient descent learning 8.5 Effect of the number of radial basis functions
8.6 Effect of the initialization of gradient descent learning 8.7 Benchmarking reformulated RBF neural networks 9. Conclusions
References