Applications of Algorithmic Information Theory

Dr. Marcus Hutter (RSISE, ANU)

MSI Computational Mathematics Seminar Series

DATE: 2009-05-11
TIME: 11:00:00 - 12:00:00
LOCATION: JD G35
CONTACT: JavaScript must be enabled to display this email address.

ABSTRACT:
Algorithmic information theory has a wide range of applications, despite the fact that its core quantity, Kolmogorov complexity, is incomputable. Most importantly, AIT allows to quantify Occam's razor, the core scientific paradigm that "among two models that describe the data equally well, the simpler one should be preferred". This lead to universal theories of induction and action, in the field of machine learning and artificial intelligence, and practical versions like the Minimum Description Length (MDL) principle. The universal similarity metric probably spawned the greatest practical success of AIT. Approximated by standard compressors like Lempel-Ziv (zip) or bzip2 or PPMZ, it leads to the normalized compression distance, which has been used to fully automatically reconstruct language and phylogenetic trees, and many other clustering problems. AIT is has been applied in disciplines as remote as Cognitive Sciences, Biology, Physics, and Economics.

Recommended reading: Applications of Algorithmic Information Theory. Scholarpedia, 2:5 (2007) 2658 http://www.scholarpedia.org/article/Applications_of_algorithmic_information_theory
BIO:
Marcus Hutter is Associate Professor in the RSISE at the Australian National University in Canberra, Australia, and NICTA adjunct. He holds a PhD and BSc in physics and a Habilitation, MSc, and BSc in informatics. Since 2000, his research is centered around the information-theoretic foundations of inductive reasoning and reinforcement learning, which resulted in 50+ published research papers and several awards. His book "Universal Artificial Intelligence" (Springer, EATCS, 2005) develops the first sound and complete theory of AI. He also runs the Human Knowledge Compression Contest (50'000 Eur H-prize).