Syllabus, Spring 1999
Psy 5038: Introduction to Neural Networks
Psychology Department, University of Minnesota
http://vision.psych.umn.edu/www/kersten-lab/courses/Psy5038/5038_99.html

 Instructor: Daniel Kersten 211 Elliott Hall 625-2589  TA: Cindee Madison, 517 Elliott Hall, Phone: 6-1337
 Office Hour: Tuesday 12:30-1:30 or by appointment  Office Hour: Thursday 10:00-11:00 or by appointment
 email: kersten@tc.umn.edu  email: cindee@eye.psych.umn.edu

Introduction to Neural Networks. (4 cr; prereq Math 3261, Psy 3031 or 5061, or #)
Introduction to large scale parallel distributed processing models in neural and cognitive science. Topics include: linear models, Hebbian rules, self-organization, non-linear models, information optimization, and representation of neural information. Applications to sensory processing, perception, learning, and memory.

Text: Introduction to Neural Networks, James Anderson, MIT Press. 1995.
Lecture notes are in Mathematica Notebook format. You can either download the Mathematica files to view (with MathReader 3.0, which is free) or to run (with Mathematica 3.0, which is not free). If you have the Adobe Acrobat plug-in, your can read and print pdf versions of the files directly with your browser. The lecture note file will be updated as the quarter progresses. If you can't wait, you can get a preview by going to last year's lecture notes.

Macintosh diskette - (3 1/2" HD) for back-ups available at Williamson Bookstore or Kinko's.
Class meeting time: Tuesdays & Thursdays, 11:15 - 12:30
Place: First week changed from original class scheduled to Eddy Hall Annex Computer Lab,
Room 62.

There will be four Mathematica based computer lab assignments, a midterm and a final examination. The grade weights are: 30% for the midterm, 30% for the final and 40% for the problem sets. The computer lab assignment solutions will be dropped off electronically.

Mathematica is available on the PowerMacs in Lind 26 and in the Eddy Hall Annex Computer Lab.

Outline

Week 1: Mar. 30 & Apr. 1. Background: neurons & models

PS 1. Introduction to Mathematica , vectors, cross-correlation - Due Tuesday, April 13
(pdf file)

Week 2: Apr. 6 & 8. Simple neural systems

Week 3: Apr. 13 & 15. Learning & memory


Week 4: Apr. 20 & 22. Non-linear networks

Week 5: Apr. 27. Gradient descent algorithms

Mid-term exam: Thursday, April 29 (Study guide)


Week 6: May 4 & 6. Representation of information


Week 7: May 11 & 13. Energy models


Week 8: May 18 & 20. Neurocomputing as Bayesian inference


Week 9: May 25 & 27. Self-organization & efficient representation


Week 10: June 1. Examples of associative computation

----
Final exam: In Class on Thursday,June 3 (Study guide)

----

Supplementary Reading

Bishop, C. M. (1995). Neural Networks for Pattern Recognition. Oxford: Oxford Univeristy Press.
Blachman, N. (1992). Mathematica: A Practical Approach . Englewood Cliffs, NJ: Prentice Hall.
Freeman, J. A. (1994). Simulating Neural Networks with Mathematica . Reading, MA: Addison-Wesley Publishing Company.
Hertz, J., Krogh, A., &;Palmer, R. G. (1991). Introduction to the theory of neural computation (Santa Fe Institute Studies in the Sciences of Complexity ed.). Reading, MA: Addison-Wesley Publishing Company.
Jordan, M. I. and Bishop. C. MIT Artificial Intelligence Lab Memo 1562, March 1996. Neural networks.
Ripley, B. D. (1996). Pattern Recognition and Neural Networks . Cambridge, UK: Cambridge University Press.
Wolfram, S. (1991). Mathematica: A System for Doing Mathematics by Computer (2nd ed.). Redwood City, CA: Addison-Wesley Publishing Company, Inc.


World Wide Web

For more information, see the Wolfram Research, Inc. Mathematica Web site, URL:
http://www.wri.com/
Once there, you can access the MathSource Electronic Library
MathSource' contains the following items relevant to neural networks:
0204-523: Exploring Neural Networks with Mathematica -- First Series
0205-906: Simulating Neural Networks with Mathematica---Electronic Supplement