The ability of a computer to learn by interacting with the environment is a fascinating subject. There are a few theoretical models of the learning process. Among them, the exact learning model represents situations in which a learner actively collects information. Specifically, the learner who tries to learn some target function f is allowed to make two kinds of queries regarding f:
Lecture 1: | Introducing the exact learning model. |
Lecture 2: | Comparison of Learning Models, and Learning Monotone DNF. |
Lecture 3: | Learning Deterministic Automata. |
Lecture 4: | Learning Deterministic Automata without Reset. |
Lecture 5: | Learning Multiplicity Automata. |
Lecture 6: | Using Multiplicity Automata to Learn Polynomials, Decision Trees, and Disjoint DNF. |
Lecture 7: | Learning read-one formulae. |
Lecture 8: | The monotone theory. |
Lecture 9: | Learning boxes in high dimension. |
Lecture 10: | The composition theorem - learning geometric objects in constant number of dimensions. |
Lecture 11: | The on-line model, learning decision lists, and the winnow algorithm. |
Lecture 12: | Efficient learning with virtual threshold gates. |
Sample file - a skeleton file for using the include file
Latex file of Lecture 1 - The latex file of the first lecture (figures omitted).
Lecture hours: | Tuesday 10:00-12:00, Building 28, Room 103, |
Reception hours: | Thursday 12:30-14:30, Building 58 (Math), Room 205. |
E-mail: | beimel at cs.bgu.ac.il |
Phone: | 647 7858 |
Course home page: | http://www.cs.bgu.ac.il/~beimel/Courses/learning.html |