Instructor:
Stephen ScottCourse home page: http://cse.unl.edu/~sscott/teach/Classes/cse978S06/
Avery 364
sscott AT cse
http://www.cse.unl.edu/~sscott/
PREREQUISITES: CSCE 310 (or equivalent programming background) and background in calculus, linear algebra, and probability and statistics. CSCE 478/878 (Machine Learning) or CSCE 970 (Pattern Recognition) is useful, but not required.
TIME: 9:30-10:45 Tuesday, Thursday
CLASSROOM: Avery 112
CSCE COURSE TRACK CLASSIFICATION: Applications track
CREDITS: 3 hours
TEXTBOOKS:
Required: Learning with Kernels by Bernhard Schölkopf and Alexander J. Smola, MIT Press, 2002.
Optional: Convex Optimization by Stephen Boyd and Lieven Vandenberghe, Cambridge University Press, 2004.
Course Description:
Building machines that learn from experience is an important research goal of artificial intelligence (AI). The field of machine learning is a subarea of AI that is concerned with the question of how to construct computer programs that automatically improve with experience. Recently, a new machine learning technique called support vector machines (SVMs) has emerged as a state-of-the-art approach. This technique is very general and powerful and has been successfully applied to problems in biological sequence analysis, text classification, image processing, data mining, and several other areas.
The goal of this course is to present the key algorithms and theory that form the core of support vector machines, including the notion of the margin, the design and use of kernels, and the formulation of a learning problem as an optimization problem that can be solved optimally. In this course we will review these basic SVM concepts as well as a few advanced topics such as kernel principal component analysis, one-class classification, regression estimation, and the appropriateness of various kernels for different applications. At the end of the course, the students will sufficiently understand the fundamentals of SVMs to design and use their own SVM approaches to various problems and to perform basic research in SVMs.
Grades in this course will be based on homework exercises (both theoretical and implementation-based), written topic summaries, and a project.
Last modified 16 August 2011.