Title: Universal Artificial Intelligence: Mathematical and Philosophical Foundations
Time: February 28 - March 3, 2006
Location: Lecture room C222
Exactum, University of Helsinki, Gustaf Hällströminkatu 2b
Schedule: Lectures 13-16.
Grading: Active participation to the lectures and doing all other parts after the course are necessary for passing the course.
Motivation: The dream of creating artificial devices that reach or outperform human intelligence is an old one, but a computationally efficient theory of true intelligence has not been found yet, despite considerable efforts in the last 50 years. Nowadays most research is more modest, focussing on solving more narrow, specific problems, associated with only some aspects of intelligence, like playing chess or natural language translation, either as a goal in itself or as a bottom-up approach. The dual, top down approach, is to find a mathematical (not computational) definition of general intelligence. Note that the AI problem remains non-trivial even when ignoring computational aspects.
Contents: In this course I develop such an elegant mathematical parameter-free theory of an optimal reinforcement learning agent embedded in an arbitrary unknown environment that possesses essentially all aspects of rational intelligence. Most of the course is devoted to giving an introduction to the key ingredients of this theory, which are important subjects in their own right: Occam's razor; Turing machines; Kolmogorov complexity; probability theory; Solomonoff induction; Bayesian sequence prediction; agents; sequential decision theory; adaptive control theory; reinforcement learning; Levin search and extensions.
References: This course is based on the book by Marcus Hutter,
Universal Artificial Intelligence, EATCS, Springer, 2004.
http://www.idsia.ch/~marcus/ai/uaibook.htm
Homework task is available from page
http://www.cs.helsinki.fi/teemu.roos/uaicourse06/homework.html
Marcus Hutter received a masters degree in computer sciences in 1992 at
the Technical University in Munich, Germany. After his PhD in
theoretical particle physics in 1995, he developed algorithms in a
medical software company for 5 years. Since 2000 he has published over
35 research papers, while working as a senior researcher at the
Artificial Intelligence institute IDSIA in Lugano, Switzerland. His
current interests are centered around reinforcement learning,
algorithmic information theory and statistics, universal induction
schemes, adaptive control theory, and related areas. In his recently
published book "Universal Artificial Intelligence" (Springer, EATCS,
2004), he unifies sequential decision theory with algorithmic
information theory. Since 2003, he has also been an honorary official
lecturer at the Technical University Munich.