Ray Solomonoff

from Wikipedia, the free encyclopedia

Ray Solomonoff (born July 25, 1926 in Cleveland , † December 7, 2009 in Cambridge (Massachusetts) ) was an American mathematician.

Solomonoff was the son of Russian Jewish immigrants and worked as an electronics instructor in the United States Navy after graduating from high school in 1944 . From 1947 he studied physics at the University of Chicago with a master’s degree in 1951. He heard there with Enrico Fermi and Rudolf Carnap, among others . From 1958 he worked for a small company, the Zator Company in Cambridge (Massachusetts), which essentially consisted of himself and the founder (1954) Calvin Mooers (as well as visiting scholars such as Marvin Minsky ) and carried out assignments for the Department of Defense. He had previously worked for her part-time. From 1962 they were called Rockford Research, and when the US government cut funding for such civilian military research as part of the Vietnam protests in 1968, Solomonoff became unemployed and founded his own consultancy, Oxbridge Research, in 1970, which he essentially consisted of and has worked for ever since worked until his death.

He was a visiting scientist for 9 months at the Artificial Intelligence Laboratory of MIT , 1990/91 at the University of Saarbrücken and at the IDSIA ( Istituto Dalle Molle di Studi sull'Intelligenza Artificiale ) in Lugano.

Solomonoff was one of the first to study the field of artificial intelligence in the 1950s . He met Marvin Minsky and John McCarthy in 1952 and was one of ten participants in the founding conference on the subject at Dartmouth College in 1956. At the conference he lectured on stochastic machine learning and published in 1957 as one of the first on the subject. His program learned arithmetic formulas through examples without invasive supervision. He also wrote some of the earliest papers on statistical analysis of networks with Anatol Rapoport (1950-1952). In the late 1950s he developed stochastic grammars and languages, and around 1960 he discovered algorithmic information theory (independent of Andrei Kolmogorow in 1965, after whom it is named as Kolmogorow Complexity ). In these works he presented both his concept of algorithmic probability and his universal theory of inductive inference.

In 2003 he was the first to receive the Kolmogorov Award from The Computer Learning Research Center at Royal Holloway College, University of London, where he was visiting professor.


  • Two kinds of probabilistic induction, Computer Journal, Volume 42, 1999, p. 256, pdf

Web links

Individual evidence

  1. Solomonoff, An Inductive Inference Machine, IRE Convention Record, Section on Information Theory, Part 2, 1957, pp. 56-62 (IEEE Symposium on Information Theory). His contribution to the Dartmouth Conference 1956: An inductive inference machine , pdf
  2. ^ Solomonoff, A preliminary report on a general theory of inductive inference, Tech. Rept. ZTB-138, Zator Company, Cambridge, Mass., November 1960. pdf . He also lectured on this at a conference at Caltech in 1960.
  3. More fully presented in: Solomonoff, A theory of inductive inference, 2 parts, Information and Control, Volume 7, 1964, pp. 1-22, 224-254
  4. Marcus Hutter, Shane Legg, Paul Vitanyi, Algorithmic Probability, Scholarpedia