EE 371D: Introduction to Neural Networks
Fall 2005




Class times: TTh: 9:30-11am, ENS 109, Unique No. 16220
Course URL: http://www.lans.ece.utexas.edu/course/ee371d/05f/index.html
Instructor: Joydeep Ghosh.
Contact: ACES 3.118; 471-8980; ghosh@ece.utexas.edu; http://www.lans.ece.utexas.edu/~ghosh
Office Hrs: M 3-4:30 pm; Th: 3:30-5pm. Other times by appointment only.
TA: TBD, ACES 3.106, @lans.ece.utexas.edu


PREREQUISITES: EE351K (Probability and Stats) or equivalent. Knowledge of Matlab will be helpful. M 340L is helpful but not needed.

This is an introductory course on neural networks and associated adaptive/learning systems. The primary emphasis is on the theory, modeling/analysis and representative applications of artificial neural networks rather than their neurophysiological plausibility. We shall look at the capabilities and limitations of several popular neural networks, mostly from the viewpoint of pattern recognition operations such as classification, regression and clustering. A nice public-domain, MATLAB based software called Netlab follows the flow of the textbook, and the Neural Network MATLAB toolbox is available as well on the LRC machines. So we shall use several MATLAB examples to illustrate the concepts involved.

Grading:
25 pts: Final exam: Th Dec 15, 9-11 am.
25+5 pts: Term paper + presentation (deadline Dec 6th)
20 pts: Mid-term exam; Tues, Oct 25, in class
25 pts: Homework and LAB assignments

For the term paper, you are encouraged to work in pairs. A list of possible topics shall be given, but you can choose almost any topic related to Neural Networks provided you have not already used this for another term paper/thesis. Each group is expected to make a 20-25 min presentation towards the end of the course.

Text: Neural Networks for Pattern Recognition by C. Bishop, Oxford University Press, 1995. Mandatory.

In addition, copies of class notes will be available through HKN.

Other Useful Books:
A more elementary level textbook is Elements of Artificial Neural Networks by Mehrotra, Mohan and Ranka, MIT Press, 1997.
Some complementary or advanced books are listed below:
(1) Neurocomputing by J. Anderson and E. Rosenfeld, MIT Press, 1988. A collection of seminar papers with introduction/commentaries.
(2) Introduction to The Theory of Neural Computation, by Hertz, Krogh and Palmer, Addison-Wesley, 1991. Nice Supplement
(3) Neural and Adaptive Systems Jose C. Principe et al. 2001. Wiley.
(4) Neural Networks: A Comprehensive Foundation by S. Haykin, Prentice-Hall (2nd Ed.), 1997. Quite comprehensive!

Course Schedule: I'll spend the first 15 lectures covering Chapters 1-9.4 (except Chapter 7) of the text. The remaining lectures will be on unsupervised learning (including self-organization) and reinforcement learning.


Disabilities: The University of Texas at Austin provides, upon request, appropriate academic adjustments for qualified students with disabilities. For more information, contact the Office of the Dean of Students at 471-6259, 471-4241 TDD, or the College of Engineering Director of Students with Disabilities, 471-4321.

ACADEMIC DISHONESTY : Cheating is very uncivilized behavior and is to be avoided at all cost. Oral discussion about assignments is not considered cheating. Copying someone else's assignment/exam part of an assignment/exam is cheating. Allegations of Scholastic Dishonesty will be dealt with according to the procedures outlined in Appendix C, Chapter 11, of the General Information Bulletin,http://www.utexas.edu/student/registrar/catalogs/.