Jürgen Schmidhuber
Jürgen Schmidhuber ( born January 17, 1963 in Munich) is a computer scientist, artist and since 1995 co-director of the Swiss Research Institute for Artificial Intelligence IDSIA. In 1993 he completed his habilitation at the Technical University of Munich and has published numerous scientific articles in the following areas: machine learning, neural networks, Kolmogorov complexity, digital physics, robotics, Hardly complexes art and theory of beauty.
The recurrent neural networks developed in his research group learning in an efficient way, some once unlernbare task: recognition of certain context-sensitive languages , robot control in partially visible environments, music composition, aspects of language processing.
His most ambitious work is possibly the Gödel machine ( 2003) to solve any problems formalisierbarer. Using an asymptotically optimal theorem the Gödel machine overwrites any part of its software (including the theorem ), once it has found a proof that this will improve their future performance.
Schmidhuber also published work on the set of possible computable universes. His " Big programmer" implements of Konrad Zuse hypothesis (1967 ) of computable physics, against which to this day there is no physical evidence. If everything is really predictable, what is the program of our world? 1997 pointed Schmidhuber out that the simplest program calculates all the universes, not just ours. A contribution from the year 2000 to continue analyzing the set of all universes with limit- computable probabilities as well as the limits of formal writability.
This work led him to generalizations of Kolmogorov complexity K ( x) of a bit string x. K ( x) is the length of the shortest program x is calculated and holds. Schmidhuber's non- retentive but converging programs are even shorter, namely the shortest possible formal descriptions Represent lead to non- enumerablen but limes computable probability measures and so-called super omegas, which are generalizations of Gregory Chaitin's " number of all mathematical wisdom " Omega is. All of this has consequences for the problem of optimal inductive inference, ie, the optimal future forecast from previously observed data.