Abstract
We exhibit upper bounds for the Vapnik-Chervonenkis (VC) dimension of a wide family of concept classes that are defined by algorithms using analytic Pfaffian functions. We give upper bounds on the VC dimension of concept classes in which the membership test for whether an input belongs to a concept in the class can be performed either by a computation tree or by a circuit with sign gates containing Pfaffian functions as operators. These new bounds are polynomial both in the height of the tree and in the depth of the circuit. As consequence we obtain polynomial VC dimension not also for classes of concepts whose membership test can be defined by polynomial time algorithms but also for those defined by well-parallelizable sequential exponential time algorithms.
Similar content being viewed by others
References
Aldaz, M., Heintz, J., Matera, G., Montaña, J.L., Pardo, L.M.: Combinatorial hardness proofs for polynomial evaluation (extended abstract). In: Mathematical Foundations of Computer science, 1998 (Brno). Lecture Notes in Computer Science, vol. 1450, pp. 167–175. Springer, Berlin (1998)
Ben-Or, M.: Lower bounds for algebraic computation trees. In: Proc. of the 15th Annual ACM Symp. on Theor. Comput., pp. 80–86. ACM, New York (1983)
Blumer, A., Ehrenfeucht, A., Haussler, D., Warmuth, M.K.: Learnability and the Vapnik-Chervonenkis dimension. J. ACM 36, 929–965 (1989)
Gabrielov, A.N., Vorobjov, N.: Complexity of computations with Pfaffian and Noetherian functions. In: Normal Forms, Bifurcations and Finiteness Problems in Differential Equations. Kluwer, Dordrecht (2004)
Gabrielov, A., Vorobjov, N., Zell, T.: Betti numbers of semialgebraic and sub-Pfaffian sets. J. Lond. Math. Soc. 69, 27–43 (2003)
von zur Gathen, J.: Parallel arithmetic computations, a survey. In: Proc. Math. Found. Comput. Sci. 13th. Lecture Notes in Computer Science, vol. 233, pp. 93–112. Springer, Berlin (1986)
Goldberg, P., Jerrum, M.: Bounding the Vapnik-Chervonenkis dimension of concept classes parametrizes by real numbers. Mach. Learn. 18, 131–148 (1995)
Grigor’ev, D., Vorobjov, N.: Complexity lower bounds for computation trees with elementary transcendental function gates. Theor. Comput. Sci. 157, 185–214 (1996)
Karpinski, M., Macintyre, A.: Polynomial bounds for VC dimension of sigmoidal and general Pffafian neural networks. J. Comput. Syst. Sci. 54, 169–176 (1997)
Khovanskii, A.: On a class of systems of transcendental equations. Sov. Math. Dokl. 22, 762–765 (1980)
Khovanskii, A.: Fewnomials. In: AMS Transl. Math. Monographs, vol. 88. AMS, Providence (1991)
Milnor, J.: On the Betti numbers of real varieties. Proc. Am. Math. Soc. 15, 275–280 (1964)
Montaña, J.L., Pardo, L.M.: Lower bounds for arithmetic networks. Appl. Algebra Eng. Commun. Comput. 4, 1–24 (1993)
Montaña, J.L., Pardo, L.M., Callau, M.: VC dimension bounds for analytic algebraic computations. In: Proc. Computing and Combinatorics, 14 Annual Intenational Conference. Lecture Notes in Computer Science, vol. 5092, pp. 62–71. Springer, Berlin (2008)
Oleinik, O.A.: Estimates of the Betti numbers of real algebraic hypersurfaces. Mat. Sb. N.S. 28, 635–640 (1951) (in Russian)
Oleinik, O.A., Petrovsky, I.B.: On the topology of real algebraic surfaces. Izv. Akad. Nauk SSSR (in Trans. Am. Math. Soc.) 1, 399–417 (1962)
Thom, R.: Sur l’Homologie des Variétés Algébriques Réelles. In: Differential and Combinatorial Topology (A Symposium in Honor of Marston Morse), pp. 255–265. Princeton University Press, Princeton (1965)
Valiant, L.G.: A theory of the learneable. Commun. ACM 27, 1134–1142 (1984)
Vapnik, V., Chervonenkis, A.: On the uniform convergence of relative frequencies of events to their probabilities. Theory Probab. Appl. 16, 264–280 (1971)
Warren, H.E.: Lower bounds for approximation by non linear manifolds. Trans. Am. Math. Soc. 133, 167–178 (1968)
Author information
Authors and Affiliations
Corresponding author
Additional information
Partially supported by MTM2007–62799 and TIN2007–67466–C02–02.
Rights and permissions
About this article
Cite this article
Montaña, J.L., Pardo, L.M. On the Vapnik-Chervonenkis dimension of computer programs which use transcendental elementary operations. Ann Math Artif Intell 56, 371–388 (2009). https://doi.org/10.1007/s10472-009-9148-3
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10472-009-9148-3
Keywords
- Concept learning
- Vapnik–Chervonenkis dimension
- Khovanskii bounds
- Pfaffian functions
- Parallel computation
- Formula size