Buch, Englisch, 145 Seiten, Paperback, Format (B × H): 155 mm x 235 mm, Gewicht: 254 g
Fast and Reliable Training Methods for Multi-Layer Perceptrons
Buch, Englisch, 145 Seiten, Paperback, Format (B × H): 155 mm x 235 mm, Gewicht: 254 g
Reihe: Perspectives in Neural Computing
ISBN: 978-3-540-76100-6
Verlag: Springer
Zielgruppe
Research
Autoren/Hrsg.
Weitere Infos & Material
1 Multi-Layer Perceptron Training.- 1.1 Introduction to MLPs.- 1.2 Error Surfaces and Local Minima.- 1.3 Backpropagation.- 2 Classical Optimisation.- 2.1 Introduction to Classical Methods.- 2.2 General Numerical Considerations.- 3 Second-Order Optimisation Methods.- 3.1 Line-Search Strategies.- 3.2 Model-Trust Region Strategies.- 3.3 Multivariate Methods for General Nonlinear Optimisation.- 3.4 Special Methods for Nonlinear Least Squares.- 3.5 Comparison of Methods.- 4 Second-Order Training Methods for MLPs.- 4.1 The Calculation of Second Derivatives.- 4.2 Reducing Storage and Computational Costs.- 4.3 Second-Order On-Line Training.- 4.4 Conclusion.- 5 An Experimental Comparison of MLP Training Methods.- 5.1 Benchmark Training Tasks.- 5.2 Initial Training Conditions.- 5.3 Experimental Results.- 6 Global Optimisation.- 6.1 Introduction to Global Methods.- 6.2 Expanded Range Approximation (ERA).- 6.3 The TRUST Method.