CPSC 636. Spring 2010

CPSC 636: Neural Networks

Spring 2010

Instructor: Ricardo Gutierrez-Osuna
Office: 520A HRBB
Phone: 979.845.2942
Email: rgutier[at]cs.tamu.edu

Course Description: The objective of this course is to familiarize students with the fundamentals of neural networksthrough selected readings from the literature.

Course organization: The course will be organized as a seminar, in which students prepare written critiques and oral presentations of selected papers, and also work in a semester-long open-ended project. Grading will be weighted as follows:

Written summaries (30%)
Oral presentations (20%)
Term project (40%)
Participation (10%)

Required background: Students are expected to have a background in signals and systems, linear algebra, calculus and probability theory.

Useful links:

How to Read an Engineering Research Paper, by William G. Griswold
Paper review guidelines, by Miryung Kim
How NOT to review a paper, by Graham Cormode
Giving a talk, by Frank Kschischang
Some rules for making a presentation, by Mihai Budiu
How to Give a Bad Talk, by David A. Patterson
Presentation Zen tips, by Garr Reynolds (great tips for graphic design)

Reading and presentation schedule (Tentative)

Date	Paper title	Presenter
01/20	Course introduction
01/22	Primer on statistical pattern recognition
01/25	Linear discriminant functions
01/27	Multilayer perceptrons
02/01	A brief history of connectionism	Jonathan
02/03	Artificial neural networks: a tutorial	Negin
02/08	Connectionist learning procedures	Siamak
02/10	Neural nets for adaptive filtering and adaptive pattern recognition	Brian
02/15	Parallel networks that learn to pronounce English text	Qiong
02/17	Nonlinear principal component analysis using autoassociative neural networks	Rhema
02/22	The cascade-correlation learning architecture	Jeremy
02/24	Class cancelled (weather)
03/01	Radial basis functions
03/03	Regularization algorithms for learning that are equivalent to multilayer networks (2)	Jin
03/08	Probabilistic neural networks	Krishna
03/10	An interactive approach for CBIR using a network of radial basis functions	Joseph
03/15	Spring break
03/17	Spring break
03/22	Proposal presentations	Everyone
03/24	Real-time neuroevolution in the NERO video game	Brittany
03/29	Competitive learning
03/31	Data exploration using self-organizing maps	Madhumanti
04/05	GTM: the generative topographic mapping	Folami
04/07	Adaptive mixtures of local experts	Neguin
04/12	Multialternative decision field theory: a dynamic connectionist model of decision making	Brittany
04/14	Neural network models of categorical perception	Jin
04/19	Finding structure in time	Jonathan
04/21	The Hopfield model	Madhumanti
04/26	Harnessing nonlinearity: predicting chaotic systems and saving energy in wireless communication	Siamak
04/28	Real-time computing without stable states: a new framework for neural computation based on perturbations	Quiong
05/03	Spiking neuron networks: a survey	Jeremy
05/05	Reading day (no class)
05/10	Final presentations (8:00-10:00AM)