VC Dimension Bounds for Analytic Algebraic Computations

  • José Luis Montaña
  • Luis Miguel Pardo
  • Mar Callau
Part of the Lecture Notes in Computer Science book series (LNCS, volume 5092)


We study the Vapnik-Chervonenkis dimension of concept classes that are defined by computer programs using analytic algebraic functionals (Nash operators) as primitives. Such bounds are of interest in learning theory because of the fundamental role the Vapnik-Chervonenkis dimension plays in characterizing the sample complexity required to learn concept classes. We strengthen previous results by Goldberg and Jerrum giving 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 an algebraic computation tree or by an algebraic circuit containing analytic algebraic gates. These new bounds are polynomial both in the height of the tree and in the depth of the circuit. This means in particular that VC dimension of computer programs using Nash operators is polynomial not only in the sequential complexity but also in the parallel complexity what ensures polynomial VC dimension for classes of concepts whose membership test can be defined by well-parallelizable sequential exponential time algorithms using analytic algebraic operators.


Concept Class Algebraic Computation Sign Node Computation Node Parallel Time 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Alonso, C.L., Montaña, J.L.: VapnikChervonenkis Dimension of Parallel Arithmetic Computations. In: Proc. of Algorithmic Learning Theory, 18th International Conference, pp. 107–119 (2007)Google Scholar
  2. 2.
    Ben-Or, M.: Lower Bounds for Algebraic Computation Trees. In: Proc. of STOC 1983, pp. 80–86 (1983)Google Scholar
  3. 3.
    Blumer, A., Ehrenfeucht, A., Haussler, A., Warmuth, M.K.: Learnability and the Vapnik-Chervonenkis Dimension. Journal of the Association for Computing Machinery 36(4), 929–965 (1989)MathSciNetzbMATHGoogle Scholar
  4. 4.
    Bochnak, J., Coste, M., Roy, M.-F.: Géométrie algébrique réelle (French) [Real algebraic geometry]. In: Ergebnisse der Mathematik und ihrer Grenzgebiete (3) [Results in Mathematics and Related Areas (3)], Berlin, vol. 12 (1987)Google Scholar
  5. 5.
    von zur Gathen, J.: Parallel Arithmetic Computations, a Survey. Math. Found. Comput. Sci. 13th Proc. (1986)Google Scholar
  6. 6.
    Goldberg, P., Jerrum, M.: Bounding the Vapnik-Chervonenkis Dimension of Concept Classes Parametrized by Real Numbers. Machine Learning 18, 131–148 (1995)zbMATHGoogle Scholar
  7. 7.
    Karpinski, M., Macintyre, A.: Polynomial Bounds for VC Dimension of Sigmoidal and General Pffafian Neural Networks. Journal of Comput. System Sci. 54, 169–176 (1997)CrossRefMathSciNetzbMATHGoogle Scholar
  8. 8.
    Milnor, J.: On the Betti Numbers of Real Varieties. Proceedings of the American Mathematical Society 15, 275–280 (1964)CrossRefMathSciNetzbMATHGoogle Scholar
  9. 9.
    Montaña, J.L., Pardo, L.M.: Lower Bounds for Arithmetic Networks. Applicable Algebra in Engineering, Communication and Computing 4(1), 1–24 (1993)CrossRefMathSciNetzbMATHGoogle Scholar
  10. 10.
    Ramanakoraisina, R.: Bezout Theorem for Nash Functions. Comm. Algebra 17(6), 1395–1406 (1989)CrossRefMathSciNetzbMATHGoogle Scholar
  11. 11.
    Valiant, L.G.: A Theory of the Learnable. Communications of the ACM 27, 1134–1142 (1984)CrossRefzbMATHGoogle Scholar
  12. 12.
    Vapnik, V., Chervonenkis, A.: On the Uniform Convergence of Relative Frequencies of Events to their Probabilities. Theory of Probability and its applications 16, 264–280 (1971)CrossRefzbMATHGoogle Scholar
  13. 13.
    Warren, H.E.: Lower Bounds for Approximation by non Linear Manifolds. Trans. A.M.S. 133, 167–178 (1968)CrossRefMathSciNetzbMATHGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2008

Authors and Affiliations

  • José Luis Montaña
    • 1
  • Luis Miguel Pardo
    • 1
  • Mar Callau
    • 1
  1. 1.Departamento de Matemáticas, Estadística y ComputaciónUniversidad de Cantabria 

Personalised recommendations