MLPfit : a simple and powerful tool for Multi-Layer Perceptrons

21/01/00


Cliquez ici pour démarrer


Table des matières

MLPfit : a simple and powerful tool for Multi-Layer Perceptrons

Acknowledgments

Part one: Multi-Layer Perceptrons

The Multi-Layer Perceptron...

Multi-Layer Perceptrons...

Applications

« Training » the neural network

The standard minimization method

Properties

A toy example

Unconstrained minimization methods (1)

Steepest descent (Cauchy 1847)

Quadratic methods

Example on the toy problem

The « hybrid » method (1)

Which method for which problem ?

Summary of part 1

Part two: The MLPfit package (1)

MLPfit 1.33

Precision

Speed/Memory

User interfaces

Code organization

Platforms

« Text » interface

Code interface

The LabVIEW interface

The PAW interface

PAW: classification

PAW: Function approximation

The Windows interface

Part three: Examples of applications

Some examples

Fit of a 1d histogram

Correction of the position measurement in the ATLAS electromagnetic calorimeter

WW -> 4q in ALEPH

Hand-written digits recognition

References

Auteur:wincenter