Polynomial Bounds for VC Dimension of Sigmoidal and General Pfaffian Neural Networks

Journal of Computer and System Sciences

Volumn 54, Issue 1, 1997, Pages 169-176

  Karpinski, Marek ^a   Macintyre, Angus ^b  

^a UNIVERSITY OF BONN   (Germany)

^b UNIVERSITY OF OXFORD   (United Kingdom)

Author keywords

[No Author keywords available]

Indexed keywords

BOOLEAN FUNCTIONS; COMPUTATIONAL COMPLEXITY; MATHEMATICAL MODELS; POLYNOMIALS;

PFAFFIAN ACTIVATION FUNCTIONS; PROGRAMMABLE PARAMETERS; SIGMOIDAL ACTIVATION FUNCTION;

NEURAL NETWORKS;

EID: 0031077292     PISSN: 00220000     EISSN: None     Source Type: Journal    
DOI: 10.1006/jcss.1997.1477     Document Type: Article

References (31)

1. M. Anthony and N. Biggs, "Computational Learning Theory: An Introduction," Cambridge Univ. Press, Cambridge, 1992.
2. M. Anthony and J. Shawa-Taylor, A result of Vapnik with applications, Discrete Appl. Math. 47 (1993), 207-217.
3. A. Borodin and P. Tiwari, On the decidability of sparse univariate polynomial interpolation, in "Proceedings, 22nd ACM STOC, 1990," pp. 535-545.
4. L. van den Dries, Tame topology and 0-minimal structures, preprint, University of Illinois, Urbana, 1992; to appear as a book.
5. L. van den Dries, A. Macintyre, and D. Marker, The elementary theory of restricted analytic fields with exponentiation, Ann. of Math. 140 (1994), 183-205.
6. P. Goldberg and M. Jerrum, Bounding the Vapnik Chervonenkis dimension of concept classes parameterized by real numbers, Mach. Learning (to appear); a preliminary version in "Proc. 6th ACM Workshop on Computational Learning Theory, 1993," pp. 361-369.
7. G. H. Hardy, Properties of logarithmic-exponential functions, Proc. London Math. Soc. 10 (1912), 54-90.
8. D. Haussler, Decision theoretic generalizations of the PAC model for neural net and other learning applications, Inform. Comput. 100 (1992), 78-150.
9. J. Hertz, A. Krogh, and R. G. Palmer, "Introduction to the Theory of Neural Computation," Addison-Wesley, Reading, MA, 1991.
10. M. W. Hirsch, "Differential Topology," Springer-Verlag, New York/Berlin, 1976.
11. M. Karpinski and A. Macintyre, Polynomial bounds for VC dimension of sigmoidal neural networks, in "Proc. 27th ACM STOC, 1995," pp. 200-208.
12. M. Karpinski and A. Macintyre, Bounding VC dimension for neural networks: Progress and prospects (invited lecture), in "Proc. EuroCOLT'95," Lecture Notes in Artif. Intell., Vol. 904, pp. 337-341, Springer-Verlag, New York/Berlin, 1995.
13. M. Karpinski and T. Werther, VC dimension and uniform learnability of sparse polynomials and rational functions, SIAM J. Comput. 22 (1993), 1276-1285.
14. A. G. Khovanski, "Fewnomials," Amer. Math. Soc., Providence, RI, 1991.
15. J. Knight, A. Pillay, and C. Steinhorn, Definable sets and ordered structures II, Trans. Amer. Math. Soc. 295 (1986), 593-605.
16. P. Koiran and E. D. Sontag, Neural networks with Quadratic VC dimension, in "Advances in Neural Information Processing Systems (NIPS'95), 1995," to appear.
17. M. C. Laskowski, Vapnik-Chervonenkis classes of definable sets, J. London Math. Soc. 45 (1992), 377-384.
18. W. Maass, Perspectives of current research about the complexity of learning on neural nets, in "Theoretical Advances in Neural Computation and Learning" (V. P. Roychowdhury, K. Y. Siu, A. Orlitsky, Eds.), pp. 295-336, Kluwer Academic, Amsterdam, 1994.
19. W. Maass, Bounds for the computational power and learning complexity of analog neural nets, in "Proc. 25th ACM STOC, 1993," pp. 335-344.
20. W. Maass, Neural nets with superlinear VC-dimension, in "Proc. of the International Conference on Artificial Neural Networks 1994, ICANN '94, Berlin, 1994," pp. 581-584; Neural Comput. 6 (1994), 875-882.
21. W. Maass, Neural nets with superlinear VC-dimension, in "Proc. of the International Conference on Artificial Neural Networks 1994, ICANN '94, Berlin, 1994," pp. 581-584; Neural Comput. 6 (1994), 875-882.
22. W. Maass, G. Schnitger, and E. D. Sontag, On the computational power of sigmoidal versus boolean threshold circuits, in "Proc. 32nd IEEE FOCS, 1991," pp. 767-776.
23. A. J. Macintyre and E. D. Sontag, Finiteness results for sigmoidal neural networks, in "Proc. 25th ACM STOC, 1993," pp. 325-334.
24. J. Milnor, On the Betti numbers of real varieties, Proc. Amer. Math. Soc. 15 (1964), 275-280.
25. J. Milnor, "Topology from the Differentiable Viewpoint," Univ. Press of Virginia, Charlottesville, 1965.
26. A. Sard, The measure of the critical points of differentiable maps, Bull. Amer. Math. Soc. 48 (1942), 883-890.
27. J. Shawe-Taylor, Sample sizes for sigmoidal neural networks, preprint, University of London, 1994; in "Proc. ACM COLT, 1995."
28. E. D. Sontag, Feedforward nets for interpolation and classification, J. Comput. System Sci. 45 (1992), 20-48.
29. G. Turan and F. Vatan, On the computation of boolean functions by analog circuits of bounded fan-in, in "Proc. 35th IEEE FOCS, 1994," pp. 553-564.
30. H. E. Warren, Lower bounds for approximation by non-linear manifolds, Trans. Amer. Math. Soc. 133 (1968), 167-178.
31. A. J. Wilkie, Model completeness results of restricted Pfaffian functions and the exponential function, J. Amer. Math. Soc., to appear.