Learning Machines
Nils J. Nilsson, 1965

Nilsson broadly defines a learning machine as "any device whose actions are influenced by past experiences". This book deals with the theory of those machines "which can be trained to recognize patterns". This includes the "PERCEPTRON and the MADELINE and MINOS networks".

The treatment of such machines is in terms of discriminant functions, and the book deals with machines having adjustable discriminant functions.

Chapter 1. Trainable Pattern Classifiers

Chapter 2. Some Important Discriminant Functions: Their Properties and their Implementations

Chapter 3. Parametric Training Methods

Chapter 4. Some Nonparametric Training Methods for Phi-Machines

Chapter 5. Training Theorems

Chapter 6. Layered Machines

Chapter 7. Piecewise Linear Machines

Appendix 1. Separation of a Quadratic Form Into Positive and Negative Parts

Appendix 2. Implementation

Appendix 3. Transformation of Normal Patterns

Top of Page | LM Opinion | Sort by Topic | Sort by Title | Sort by Author