A coding theorem for Enumerable Output Machines

Recently, Schmidhuber proposed a new concept of generalized algorithmic complexity. It allows for the description of both finite and infinite sequences. The resulting distributions are true probabilities rather than semimeasures. We clarify some points for this setting, concentrating on Enumerable O...

Full description

Saved in:
Bibliographic Details
Published inInformation processing letters Vol. 91; no. 4; pp. 157 - 161
Main Author Poland, Jan
Format Journal Article
LanguageEnglish
Published Amsterdam Elsevier B.V 31.08.2004
Elsevier Science
Elsevier Sequoia S.A
Subjects
Algorithmic information theory
Algorithmics. Computability. Computer arithmetics
Algorithms
Applied sciences
Codes
Coding theorem
Computer science; control theory; systems
Enumerable Output Machine
Exact sciences and technology
Information processing
Kolmogorov complexity
Machinery
Output
Studies
Theorems
Theoretical computing
Theory of computation
Theory of computation
Enumerable Output Machine
Coding theorem
Kolmogorov complexity
Algorithmic information theory
Computer theory
Enumerable output machine
Coding
Algorithm complexity
Information processing
Algorithm analysis
Online AccessGet full text

Cover

More Information
Summary:Recently, Schmidhuber proposed a new concept of generalized algorithmic complexity. It allows for the description of both finite and infinite sequences. The resulting distributions are true probabilities rather than semimeasures. We clarify some points for this setting, concentrating on Enumerable Output Machines. As our main result, we prove a strong coding theorem (without logarithmic correction terms), which was left as an open problem. To this purpose, we introduce a more natural definition of generalized complexity.
ISSN:0020-0190
1872-6119
DOI:10.1016/j.ipl.2004.05.002