## A computer scientist’s view of life, the universe, and everything (1997)

Venue: | Foundations of Computer Science: Potential - Theory - Cognition |

Citations: | 41 - 15 self |

### BibTeX

@INPROCEEDINGS{Schmidhuber97acomputer,

author = {Jürgen Schmidhuber},

title = {A computer scientist’s view of life, the universe, and everything},

booktitle = {Foundations of Computer Science: Potential - Theory - Cognition},

year = {1997},

pages = {201--208},

publisher = {Springer}

}

### Years of Citing Articles

### OpenURL

### Abstract

Is the universe computable? If so, it may be much cheaper in terms of information requirements to compute all computable universes instead of just ours. I apply basic concepts of Kolmogorov complexity theory to the set of possible universes, and chat about perceived and true randomness, life, generalization, and learning in a given universe. Preliminaries Assumptions. A long time ago, the Great Programmer wrote a program that runs all possible universes on His Big Computer. “Possible ” means “computable”: (1) Each universe evolves on a discrete time scale. (2) Any universe’s state at a given time is describable by a finite number of bits. One of the many universes is ours, despite some who evolved in it and claim it is incomputable. Computable universes. Let TM denote an arbitrary universal Turing machine with unidirectional output tape. TM’s input and output symbols are “0”, “1”, and “, ” (comma). TM’s possible input programs can be ordered

### Citations

555 |
Three approaches to the quantitative definition of information,” Probl
- Kolmogorov
- 1965
(Show Context)
Citation Context ...t two reasons for this: 1. Shortest algorithm cannot be found. It can be shown that there is no algorithm that can generate the shortest program for computing arbitrary given data on a given computer =-=[2, 9, 1]-=-. In particular, our physicists cannot expect to find the most compact description of our universe. 2. Additional problems of the Heisenberg type. Heisenberg tells us that we cannot even observe the p... |

423 |
A formal theory of inductive inference
- Solomonoff
- 1964
(Show Context)
Citation Context ...t two reasons for this: 1. Shortest algorithm cannot be found. It can be shown that there is no algorithm that can generate the shortest program for computing arbitrary given data on a given computer =-=[2, 9, 1]-=-. In particular, our physicists cannot expect to find the most compact description of our universe. 2. Additional problems of the Heisenberg type. Heisenberg tells us that we cannot even observe the p... |

336 | Algorithmic Information Theory
- Chaitin
- 1987
(Show Context)
Citation Context ...t two reasons for this: 1. Shortest algorithm cannot be found. It can be shown that there is no algorithm that can generate the shortest program for computing arbitrary given data on a given computer =-=[2, 9, 1]-=-. In particular, our physicists cannot expect to find the most compact description of our universe. 2. Additional problems of the Heisenberg type. Heisenberg tells us that we cannot even observe the p... |

131 | The lack of a priori distinction between learning algorithms
- Wolpert
- 1996
(Show Context)
Citation Context ...s (those that can be computed by long algorithms only), successful generalization from previous experience is not possible. Neither is inductive transfer. This simple insight is related to results in =-=[10]-=-. Learning. Given the above, since learning means to use previous experience to improve future performance, learning is possible only in the few regular universes (no learning without compressibility)... |

103 |
Random conservation inequalities; information and independence in mathematical theories
- Levin
- 1984
(Show Context)
Citation Context ...ing paradigms, the entire life is considered for performance evaluations. Experiments in [7, 6] show the paradigm's practical feasibility. For instance, in [7] A includes an extension of Levin search =-=[4]-=- for generating the PMPs. Philosophy Life after death. Members of certain religious sects expect resurrection of the dead in a paradise where lions and lambs cuddle each other. There is a possible con... |

63 | Shifting inductive bias with success-story algorithm, adaptive Levin search, and incremental self-improvement
- Schmidhuber, Zhao, et al.
- 1997
(Show Context)
Citation Context ...it is setting the stage, the framework provides a basis for "learning how to learn". Unlike with previous learning paradigms, the entire life is considered for performance evaluations. Exper=-=iments in [7, 6]-=- show the paradigm's practical feasibility. For instance, in [7] A includes an extension of Levin search [4] for generating the PMPs. Philosophy Life after death. Members of certain religious sects ex... |

49 | Discovering neural nets with low kolmogorov complexity and high generalization capability - Schmidhuber - 1997 |

38 | Reinforcement Learning with Self-Modifying Policies
- Schmidhuber, Zhao, et al.
- 1997
(Show Context)
Citation Context ...it is setting the stage, the framework provides a basis for "learning how to learn". Unlike with previous learning paradigms, the entire life is considered for performance evaluations. Exper=-=iments in [7, 6]-=- show the paradigm's practical feasibility. For instance, in [7] A includes an extension of Levin search [4] for generating the PMPs. Philosophy Life after death. Members of certain religious sects ex... |

33 |
A mathematical theory of communication, part i
- Shannon
- 1948
(Show Context)
Citation Context ...ompressible. How? Well, there is a comparatively short algorithm that simply codes probable next states by few bits, and unlikely next states by many bits, as suggested by standard information theory =-=[8]-=-. More regularity than we think? The longer the shortest program computing a given universe, the more random it is. To certain observers, certain universes appear partly random although they aren't. T... |

23 |
of information (nongrowth) and aspects of the foundation of probability theory
- Levin
(Show Context)
Citation Context ...o the right, do: (1) Append a new square. (2) With probability 1 3 fill it with a "0"; with probability 1 3 fill it with a "1"; with probability 1 3 fill it with a ",". P=-=rograms are "self-delimiting" [3, 1]-=- --- once U halts due to computations based on the randomly chosen symbols (the program) on its input tape, there won't be any additional program symbols. We obtain PU (s) = X p:U computes s from p an... |