Courses
  /  
Descriptions
EECS 435: Neural Networks

Quarter Offered

Fall : 2-2:50 MWF ; Lin

Prerequisites

EECS 205-EA1 (linear algebra and MATLAB) or equivalent, Math 214 (calculus) or equivalent, EECS 302 (probability) or equivalent

Description

OBJECTIVES: The objective of this course is to provide students with a basic understanding of the theoretical foundations and applications of artificial neural networks.

COURSE DIRECTOR: Prof. Wei-Chung Lin

RECOMMENDED TEXTBOOK:

  • M. T. Hagan, H. B. Demuth, M. H. Beale, O. D. Jesus, Neural Network Design, 2nd Edition.   A PDF version of the textbook can be downloaded FREE from  http://hagan.okstate.edu/NNDesign.pdf

COURSE CONTENTS:

  • Week 1 (9/21-9/25)

Introduction to artificial neural networks (Ch. 1), Neuron model and network architectures (Chs. 2, 3), Perceptron learning rule (Ch. 4).

  • Week 2 (9/28-10/2)

Perceptron learning rule (continued), Linear Transformations for Neural Networks (Chs. 5, 6),

  • Week 3 (10/5-10/9)

Linear Transformations for Neural Networks (continued), Supervised Hebbian learning (Ch. 7), Optimal linear associative memories, Performance surfaces and optimum points (Ch. 8) 

  • Week 4 (10/12-10/16)

 Performance Optimization (Ch. 9), Widrow-Hoff learning (Ch. 10)

  • Week 5 (10/19-10/23)

Widrow-Hoff learning (continued)

  • Week 6 (10/26-10/30)

Midterm exam (10/26, Monday), Principal-components analysis

  • Week 7 (11/2-11/6)

Backpropagation learning algorithms (Ch. 11)

  • Week 8 (11/9-11/13)

Variations on Backpropagation (Ch. 12), Radial Basis networks (Ch. 16)

  • Week 9 (11/16-11/20)

Stability, Continuous Hopfield networks

  • Week 10 (11/23-11/27)

 Continuous Hopfield networks (continued), Discrete Hopfield networks

  • Week 11 (11/30-12/4)

Competitive networks (Ch. 15), Counterpropagation networks

GRADING:

  • Homework Assignments:  50%
  • Midterm exam: 20% (in class, October 26, Monday)
  • Final exam: 30%