CmpE 545 Artificial Neural Networks

Spring Semester

Ethem ALPAYDIN

Textbook

Introduction to Machine Learning, The MIT Press, 2004. One copy is on reserve at the Library.

Topics

  1. Chap 10. Linear Discrimination
  2. Chap 11. Multilayer Perceptrons
  3. Chap 12. Local Models: Radial Basis Functions and Mixture of Experts
  4. Chap 14. Assessing and Comparing Classification Algorithms
  5. Chap 15. Combining Multiple Learners
  6. Chap 16. Reinforcement Learning
  7. Advanced topics on Bayesian Networks, Kernel methods, and Nonlinear dimensionality reduction

Goals

An artificial neural network is an interconnected set of simple concurrently operating processing units. This course follows CmpE 544 Pattern Recognition and dicusses novel methods using such networks in a comparitive manner with the classical approaches discussed in CmpE 544.

Prerequisite by Topic

CmpE 544, Undergraduate level calculus, probability theory. Prior experience in a high-level programming language.

Computer Usage

Almost all homeworks require computer simulations. I recommend using Matlab.

Grading

  • 1 Project 0.30
  • 1 Final 0.30
  • Homeworks 0.40
    • Last modified on Oct 26, 2004.