Gives an introduction to basic (artificial) neural network architectures and learning rules. Emphasis is placed on mathematical analysis of these networks, on methods of training them, and on their application to practical problems in areas such as pattern recognition, signal processing, and control systems. The course will show how to construct neural networks and train them to perform useful functions. Neural networks are good at fitting non-linear functions and recognizing patterns. Consequently, they have wide application in the aerospace, automotive, banking, defense, electronics, entertainment, financial, insurance, manufacturing, oil and gas, robotics, telecommunications, and transportation industries. At the conclusion of the course, students will be able to understand and analyze the major types of neural networks and will be able to design and implement networks to solve practical problems.

Introduction Neuron Model and Network Architecture Illustrative Example Perceptron Learning Rule Signal and Weight Vector Spaces Linear Transformations for Neural Networks Supervised Hebb Performance Surfaces and Optimum Points Performance Optimization Widrow Hoff backpropagation Variations on Backpropagation

Develop understanding and skills for practical application of neural networks.

Linear algebra.

MATLAB and the Neural Network Toolbox for MATLAB will be used. The student version of MATLAB can be purchased from the CU Book Store. The student version of the Neural Network Toolbox can be purchased from MathWorks at www.mathworks.com. MATLAB and the Neural Network Toolbox for MATLAB are also available on-campus in the computing labs.