We must Choose the Simplest Physical Theory: Levin-Li-Vitányi Theorem and its Potential Physical Applications
If several physical theories are consistent with the same experimental data, which theory should we choose? Physicists often choose the simplest theory; this principle (explicitly formulated by Occam) is one of the basic principles of physical reasoning. However, until recently, this principle was mainly a heuristic because it uses the informal notion of simplicity.
With the explicit notion of simplicity coming from the Algorithmic Information theory, it is possible not only to formalize this principle in a way that is consistent with its traditional usage in physics, but also to prove this principle, or, to be more precise, deduce it from the fundamentals of mathematical statistics as the choice corresponding to the least informative prior measure. Potential physical applications of this formalization (due to Li and Vitányi) are presented.
In particular, we show that, on the qualitative level, most fundamental ideas of physics can be re-formulated as natural steps towards choosing a theory that is the simplest in the above precise sense (although on the intuitive level, it may seem that, e.g., classical physics is easier than quantum physics): in particular, we show that such ideas as Big Bang cosmology, atomism, uncertainty principle, Special Relativity, quark confinement, quantization, symmetry, supersymmetry, etc. can all be justified by this (Bayesian justified) preference for formalized simplicity.
Key wordsKolmogorov complexity Algorithmic Information theory Occam razor Bayesian statistics Fundamental physics
Unable to display preview. Download preview PDF.
- 3.A. Einstein, “On the method of theoretical physics”, The Herbert Spencer Lecture delivered at Oxford on June 10, 1933. Reprinted in: A. Einstein, Ideas and opinions, Crown Publishers, N.Y., 1954, pp. 270–276.Google Scholar
- 7.A. M. Finkelstein and V. Kreinovich, “Derivation of Einstein’s, Brans-Dicke and other equations from group considerations,” On Relativity Theory. Proceedings of the Sir Arthur Eddington Centenary Symposium, Nagpur India 1984, Vol. 2, Y. Choque-Bruhat and T. M. Karade (eds), World Scientific, Singapore, 1985, pp. 138–146.Google Scholar
- 8.A. M. Finkelstein, V. Kreinovich, and R. R. Zapatrin, “Fundamental physical equations uniquely determined by their symmetry groups,” Lecture Notes in Mathematics, Springer-Verlag, Berlin-Heidelberg-N.Y., Vol. 1214, 1986, pp. 159–170.Google Scholar
- 9.Q. Gao and M. Li, “An application of minimum description length principle to online recognition of handprinted alphanumerals”, In: Proc. 11th Int’s Join Conferences on Artificial Intelligence IJCAI, Morgan Kaufmann, San Mateo, CA, 1989, pp. 843–848.Google Scholar
- 10.Group theory in physics: proceedings of the international symposium held in honor of Prof. Marcos M o shins ky, Cocoyoc, Morelos, Mexico, 1991, American Institute of Physics, N.Y., 1992.Google Scholar
- 15.M. Li and P. M. B. Vitányi, “Computational machine learning in theory and practice”, In: J. van Leeuwen (ed.), Computer Science Today, Recent Trends and Developments, Springer Lecture Notes in Computer Science, Springer-Verlag, Berlin-Heidelberg-N.Y., 1995, Vol. 1000, pp. 518–535.Google Scholar
- 17.Ch. Misner, K. S. Thorne, and J. A. Wheeler, Gravitation, W. H. Freeman and Co., San Francisco, 1973.Google Scholar
- 19.E. P. D. Pednault, “Some experiments in applying inductive inference principles to surface reconstruction”, In: Proc. 11th International Joint Conferences on Artificial Intelligence IJCAI, Morgan Kaufmann, San Mateo, CA, 1989, pp. 1603–1609.Google Scholar
- 21.J. J. Rissanen, Stochastic complexity and statistical inquiry, World Scientific, Singapore, 1989.Google Scholar
- 23.I. Rosenthal-Schneider, “Presuppositions and anticipations”, In: P. A. Schlipp (ed.), Albert Einstein: philosopher-scientist, Tbdor Pubi., N.Y., 1951.Google Scholar
- 25.Symmetries in physics: proceedings of the international symposium held in honor of Prof. Marcos Moshinsky, Cocoyoc, Morelos, Mexico, 1991, Springer-Verlag, Berlin, N.Y., 1992.Google Scholar
- 26.P. M. B. Vitányi and M. Li, “Ideal MDL and its relation to Bayesianism”, In: D. Dowe, K. Korb, and J. Oliver (eds.), Proc. ISIS: Information, Statistics, and Induction in Science Conference, World Scientific, Singapore, 1996, pp. 282–291.Google Scholar
- 27.P. M. B. Vitányi and M. Li, Minimum description length induction, Bayesianism, and Kolmogorov complexity, Manuscript, CWI, Amsterdam, 1996.Google Scholar