CS 6673 Neural Network Computing (1472I), Spring 2009

This course gives an introduction to neural network models and their applications. Organization and learning in neural network models including perceptrons, adalines, backpropagation networks, recurrent networks, adaptive resonance theory and the neocognitron are discussed. Application in areas such as decision systems, nonlinear control, speech processing and vision is explored.

The pre-requisite for this course is CS 540 or its equivalent. Undergraduate student may take this course if they have already taken CS 1114, and the undergraduate math sequence.


Instructor :

K. Ming Leung

 

E-mail :

mleung@duke.poly.edu

 

Office :

LC 127

 

Office hours :

 

Textbook :

"Neural Network Design", by Martin T. Hagan, Howard B. Demuth, Mark H. Beale, ISBN-10: 0-9717321-0-8.

 

Classroom :

RH 317

 

Grading System (Tentative):

Attendance/Participation (5%), HW (35%), Midterm-Exam (30%), Project (30%).

 

Lecture Notes :

Introduction:

    Introduction to Artificial Neuron Networks.

    Neural Networks for Pattern Classification.

    Perceptron for Pattern Classification.

    ADALINE for Pattern Classification.

    Backpropagation for Multilayer Neural Network

    Learning Vector Quantization.

    Neural Networks for Pattern Association

    Iterative Bidirectional Memory

    Iterative Discrete Hopfield Net

    Hamming Net

    Self-Organized Map

    Adaptive Resonance Theory 1

 

Assignments :

HW01 (Due February 10)

HW02 (Due March 3)

HW03 (Due March 24): Solution

HW04 (Due March 24): Solution

HW05 (Due April 14)

HW06 (Due May 5)

HW07 (Due April 28)

 

Midterm: (open book midterm) April 7.