Course Announcement for Spring 2007
CSCE 970: Pattern Recognition
(with an emphasis on graphical models)
Instructor:
Stephen ScottPrevious offering: http://cse.unl.edu/~sscott/teach/Classes/cse970S03/
Avery 364
sscott AT cse
http://www.cse.unl.edu/~sscott/
Pattern recognition is concerned with the question of how to automatically detect subtle patterns in data. Pattern recognition has for many years been applied to numerous areas, including computer vision, signal processing, bioinformatics, data mining, and many more.
In this course you will learn several of the fundamentals and current trends in pattern recognition, with an emphasis on so-called "graphical models" such as Bayesian networks and hidden Markov models. These models are rigorously justified, provide a distributed knowledge representation, and are as understandable as a rule base. They deal particularly well with uncertainty, and they can be manually generated by consultation of an expert, or inductively built by machine learning. These approaches are used for modelling knowledge in gene regulatory networks, medicine, engineering, text analysis, image processing, data fusion, and decision support systems. A sampling of specific applications includes (1) automatic interpolation of raw data from interplanetary probes and deep space explorations; (2) modeling pilot-aircraft interactions; (3) the "Office Assistant" that was introduced in the Office 95 suite of desktop products; and (4) modeling of DNA and protein sequences.
In contrast with previous offerings of this course, there will be no homeworks or project. Instead, the course requirements will be regular attendance, a major written and oral presentation of course material, substantive edits of other students' written presentations of course material, and at the end of the term a presentation of your own research, tied as much as possible to the course content. For information on the previous offering of this course, see http://cse.unl.edu/~sscott/teach/Classes/cse970S03