Computational Theories of Learning and
Meeting Times and Locations:
Lecture: 0.75 unit, Tue/Thu 9:30 - 10:45,
Professor: Dan Roth
Office: 2101 DCL
Office Hours: Thursday 1:30-3pm
Phone: (217) 244-7068
The purpose of the course is to acquaint students with the theoretical
foundations of machine learning and intelligent inference.
Along with providing an introduction to this field the emphasis will
be on providing familiarity with some topics in current research.
The focus, beyond introducing the main computational models, would be two fold:
- The study of learning algorithms that
are both amenable to mathematical analysis and make sense empirically
(in terms of performance and scalability) and
- Ways to integrate theories of
learning with those of reasoning,
- Here is a
tentative plan of the course.
The course is targeted at graduates and advanced
undergraduates. Ideally, students should have background in basic
computation theory and algorithms, basic combinatorics and
probability, and introductory AI.
The course will not have any exam. Instead there will be several other
- Scribe notes: Each lecture one of the students will be assigned
the job of "scribing" the lecture, for later distribution to the
class. The notes should not be simple copy of the what is written on
the board. It has to be written so that it reflects understanding of
A latex form as well as draft scribe from last year
(for some of the lectures) will be available
- Presentation: All the students will present a lecture in
class. The material for the lecture will be from a research paper.
A preliminary list of papers is available.
- All students will write a short (half a page) review of the target paper.
- In addition, there may be 1-3 non-obligatory problem sets on the
introductory material. The goal is to help you make sure you
understand the material.
The articles will be distributed in class. Lecture notes and handouts
will be available from the course home page http://L2R.cs.uiuc.edu/~danr/Teaching/CS397-99/