1 / 21

Physically-Relativized Church-Turing Hypotheses

Physically-Relativized Church-Turing Hypotheses. Martin Ziegler Theoretical Computer Science University of Paderborn 33095 GERMANY. „Does there exist a physical system of computational power strictly exceeding that of a Turing machine?“. for example able to solve the Halting problem ?

akando
Download Presentation

Physically-Relativized Church-Turing Hypotheses

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Physically-Relativized Church-Turing Hypotheses Martin Ziegler Theoretical Computer Science University of Paderborn 33095 GERMANY

  2. „Does there exist a physical system of computational power strictly exceeding that of a Turing machine?“ • for example able to solve the Halting problem? • or in polynomial time some NP-complete problem? • Answer has tremendous effects on our conception of • nature (universe as a computer?, cf. eg. Seth Lloyd) • Turing machines: universal model of computation • in computer science (WHILE-programs, λ-calculus) • in mathematics (μ-recursive function class) • and in physics? • engineering actual computing devices (Intel, AMD)

  3. „Does there exist a physical system of computational power strictly exceeding that of a Turing machine?“ • (Physical/strong) Church-Turing Hypothesis: No! • Audience Poll: • Do you believe in this hypothesis? • Proof? • What is a physical system, anyway?

  4. Feynman, Shor, Deutsch, GroverAdamyan, Calude, Dinneen, Pavlov, Kieu

  5. Malament, Hogarth, Nemeti …

  6. A.C.-C.Yao, W.D.Smith, K.Svozil …

  7. Reif & Tate & Yoshida (1994), Oltean (2006ff), Woods

  8. So, what is a Physical System? WB2 WB1 WB3 • Beggs&Tucker (2007): „[…] we should […] use a physical theory to define precisely the class of physical systems under investigation“ Celestial Mechanics, Newtonian Mechanics, Continuum Mechanics, Magneto-statics, Electrostatics, Ray Optics, Gaussian Optics, Electrodynamics, Special Relativity, General Relativity, Quantum Mechanics, Quantum Field Theory Ludwig: „Die Grundstrukturen einer physikalischen Theorie“, Springer (1990) Schröter: „Zur Meta-Theorie der Physik“, de Gruyter (1996) • A physical theory Φ consists of 3 parts: • a mathematical theory MT • a part WBof nature it aims to describe • a correspondence AP from WB to MT „Reality“ described/covered by a patchwork of physical theories

  9. Church-Turing Hypothesis relative to a Physical Theory • Principle 2.2 in Beggs&Tucker (2007):“Classifying computers in a physical theory.” „Does there exist in Φa physical system of compu- tational power exceeding that of a Turing machine?“ „Does there exist in Φa physical system of compu- tational power exceeding that of a Turing machine?“ That is, fix a physical theory Φ and consider validity of the Church-Turing Hypothesis (CTH) relative to Φ. → Research Programme: For various physical theories Φ, investigate CTHΦ. Principle 2.3 in Beggs&Tucker (2007): Mapping the border between computer and hyper-computer in physical theory.

  10. Research Programme For various physical theories Φ, investigate CTHΦ. • For a fixed Φ, does there exist in Φ • a system able to solve the Halting problem? • or in polynomial time some NP-complete problem? Compare Baker&Gill&Solovay (1975):„Relativizations of the P=?NP Question”: For one oracle A, provably PA=NPA; for another oracle B, provably PB≠NPB. Ontological commitment; again Beggs&Tucker (2007): “It does not matter whether we think of the theory Φ as true, or roughly applicable, or know it to be false”

  11. Ontological Commitment GUT/ ToE GUT/ToE WB2 WB1 WB3 It does not matter whether we think of the theory Φ as true, or roughly applicable, or know it to be false. • Is there a theory which is not „false“ somehow? • → Grand Unified Theory/Theory of Everything • → dream, not science • Pragmatic: each Φ describes somepart of reality more or less accurately „Reality“ described/covered by a patchwork of physical theories

  12. Ontological Commitment II It does not matter whether we think of the theory Φ as true, or roughly applicable, or know it to be false. • Pragmatic: each Φ describes somepart of reality more or less accurately • Compare Models of Computationin Theoretical Computer Science: • Is a ZX81 more appropriately described by a TM or by a DFA? • Even ‘small‘ WB (=area of applica-bility) may have ‘large‘ applications! • Ohm‘s Law & CM vs. QED

  13. Computational Physics • Simulating of a (class of) physical systems Φ

  14. Computational Physics • Computational complexity of simulating a (class of) physical systems Φ: • –complete if, in Φ, there exist systems implementing, e.g., • Boolean circuit evaluation • Travelling sales tour search • Universal Turing computation P, NP, PSPACE, REC Principle 2.2 in Beggs&Tucker (2007):“Classifying computers in a physical theory.” → CTHΦ as approach to dis-/prove optimality of algorithms in computational physics!

  15. Research Programme Principle 2.3 in Beggs&Tucker (2007): Mapping the border between computer and hyper-computer in physical theory. For various physical theories Φ, investigate CTHΦ. Start with ‘simplest‘ theories! It does not matter whether we think of the theory Φ as true, or roughly applicable, or know it to be false.

  16. Example: Celestial Mechanics Research Programme: For various physical theories Φ, investigate CTHΦ. It does not matter whether we think of the theory Φ as true, or roughly applicable, or know it to be false. • Various physical theories: • full relativistic effects • Newton gravitation • Kepler ellipses around center of gravity, w/o interaction • Copernican heliocentrism • Ptolemaic geocentrism • planar, circular rotation

  17. Example: Celestial Mechanics • Various physical theories • Newton gravitation: • PSPACE-complete [Reif&Tate‘93]undecidable [W.D.Smith’06, K.Svozil‘07] • planar, circular rotation NC1…#P-compl.

  18. Example: Classical Mechanics 2H 1H 3,4H „Does there existin CMa physical system of compu- tational power exceeding that of a Turing machine?“ • An ideal solid can encode the Halting problem • and may then be used to solve it by probing:

  19. Existence in Physical Theories CM should support only solidswhich can be ‚constructed‘ (e.g. cut/carved) from few basic ones (e.g. cuboid) What makes CM unrealistic with respect to computability? 1) Real bodies are not infinitely divisible. But even if so (ontological commitment!): 2) In order to solve the Halting problem, does there exist a solid with it encoded? • When is a mathematical object considered to exist? • If one can actually construct this object. („Constructivism“) • If its non-existence raises a contradiction. (indirect proof, e.g. Markov‘s Principle) • If the hypothesis of its existence does not raise a contradiction. (e.g. Zorn‘s Lemma is consistent with ZF)

  20. Conclusion • Current hot disputes on validity of Church-Turing hypothesis • mostly due to vagueness of the underlying notion of „nature“: • ‚counterexamples‘ (=physical systems ‘solving‘ the Halting problem) exploit some physical theory Φto its limits • Better always speak of the CTH relative to a specific Φ. • independent of whether (and where) Φ is ‘realistic‘ or not. • Investigate, for various Φ, the computational power of Φ • → Lower complexity bounds in computational physics • Realistic physical theory Φ=(MT,AB,WB) should • make the Church-Turing hypothesis a theorem • (meta-principle, like gauge-invariance or energy conservation) • and employ some sort of constructivism in WB.

  21. Thanks for your attention! Heinz Nixdorf Institute & Dept of Computer Science University of Paderborn Fürstenallee 11 33095 Paderborn, Germany Tel.: +49 (0) 52 51/60 30 67 Fax: +49 (0) 52 51/62 64 82 E-Mail: ziegler@upb.de http://www.upb.de/cs/ziegler.html

More Related