CAP6610: Machine Learning

What is about this course?

Machine learning is concerned with the question of how to construct computer programs that learn in a broad sense. The immediate benefit of machine learning would be to enable AI programs to improve their performance automatically over time. For example, a chess program can improve its game plan against its opponent through playing. A robot can recognize a particular kind of object more accurately though repeated presentation of the object image. At a fundamental level, a machine with a clearly demonstrated ability to learn would answer the question of whether machines can exhibit true intelligence.

This course covers several important learning paradigms including learning from examples, decision tree learning, neural networks, support vector machines, Bayesian learning, learning rules, analytical learning, and reinforcement learning. Important learning algorithms are taught in detail and illustrated with real-world applications. In addition, computational learning theory including PAC learning, VC-dimension, and sample complexity is examined.

Recently, data mining has become an important application area of machine learning. Data mining deals with the discovery of hidden knowledge, patterns, and rules from large databases. This course also has a special series of discussions on this topic.

Who teaches the course?

LiMin Fu

Computer and Information Sciences Dept., University of Florida, Gainesville, Florida 32611
Email address: fu@cise.ufl.edu
Office: E340 CSE
Phone: (352)392-1485

When?

Class meeting on Campus: MWF/12:50 - 1:40 p.m. (Eastern Time). at MCCB G108
Office Hours: Wed, Fri/2:00 - 3:00 p.m. (Eastern Time)

Teaching Material:

Fu's Lecture Notes:

Introduction to Machine Learning

Concept Learning

Decision Tree Learning

Artificial Neural Networks

Evaluation

Bayesian Learning

Computational Learning Theory

Reinforcement Learning

Learning Rules

Genetic Algorithms

Textbook (required):

Title: "Machine Learning"
Author: Tom Mitchell
Publisher: McGraw-Hill, Inc., New York
Year: 1997
ISBN: 0-07-042807-7

Supplementary Material:

From WWW resources

Data Mining

Genetic Algorithms

Genetic Programming

Prerequisites:

In general, backgrounds in artificial intelligence and in statistics are helpful.

Course Objectives:

Understand the basic concepts
Learn the techniques
Learn the applications

Course Requirements:

Four sets of homework (15% of final grade). No late homework unless in extraordinary circumstances.
Three projects (20% of final grade).
A datamining research project with oral presentation (5% of final grade)
Mid-term (25% of final grade) (Scope: Chap 1-4): March 7 (Fri), 12:50 - 1:40 p.m.
Final-Exam (35% of final grade): (Scope: with emphasis on hypothesis evaluation, Bayesian learning, SVMs, computational learning theory, genetic algorithms, and reinforcement learning), May 1 (Thur.), 10:00 - 12:00 p.m.

Introduce Yourself:

Please enter your personal information by clicking on Register: Introduce yourself

Check Your Grade On-Line:

You may check your grades (homework, projects, exams, overall) any time as long as they are available by clicking on Check Your Grade. Notice, however, you need to have an ID name in order to use this facility.