Neural networks and complexity theory

  • Pekka Orponen
Invited Lectures
Part of the Lecture Notes in Computer Science book series (LNCS, volume 629)


We survey some of the central results in the complexity theory of neural networks, with pointers to the literature.


Boolean Function Finite Automaton Boltzmann Machine Attraction Basin Polynomial Size 
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.
    Aarts, E., Korst, J. Simulated Annealing and Boltzmann Machines. John Wiley & Sons, Chichester, 1989.zbMATHGoogle Scholar
  2. 2.
    Alon, N. Asynchronous threshold networks. Graphs and Combinatorics 1 (1985), 305–310.zbMATHMathSciNetGoogle Scholar
  3. 3.
    Alon, N., Dewdney, A. K., Ott, T. J. Efficient simulation of finite automata by neural nets. J. Assoc. Comp. Mach. 38 (1991), 495–514.zbMATHMathSciNetGoogle Scholar
  4. 4.
    Anderson, J. A., Rosenfeld, E. (eds.) Neurocomputing: Foundations of Research. The MIT Press, Cambridge, MA, 1988.Google Scholar
  5. 5.
    Anderson, J. A., Pellionisz, A., Rosenfeld, E. (eds.) Neurocomputing 2: Directions for Research. The MIT Press, Cambridge, MA, 1991.Google Scholar
  6. 6.
    Balcázar, J. L., Díaz, J., Gabarró, J. On characterizations of the class PSPACE/poly. Theoret. Comput. Sci. 52 (1987), 251–267.zbMATHMathSciNetCrossRefGoogle Scholar
  7. 7.
    Blum, A. L., Rivest, R. L. Training a 3-node neural network in NP-complete. Neural Networks 5 (1992), 117–127.CrossRefGoogle Scholar
  8. 8.
    Bruck, J. On the convergence properties of the Hopfield model. Proc. of the IEEE 78 (1990), 1579–1585.CrossRefGoogle Scholar
  9. 9.
    Cybenko, G. Approximation by superposition of a sigmoidal function. Math. of Control, Signals, and Systems 2 (1989), 303–314.zbMATHMathSciNetGoogle Scholar
  10. 10.
    Floréen, P., Orponen, P. Attraction radii in Hopfield nets are hard to compute. Manuscript submitted for publication, 7 pp., April 1992.Google Scholar
  11. 11.
    Floréen, P., Orponen, P. On the computational complexity of analyzing Hopfield nets. Complex Systems 3 (1989), 577–587.zbMATHMathSciNetGoogle Scholar
  12. 12.
    Fogelman, F., Goles, E., Weisbuch, G. Transient length in sequential iterations of threshold functions. Discr. Appl. Math. 6 (1983), 95–98.zbMATHMathSciNetCrossRefGoogle Scholar
  13. 13.
    Fogelman, F., Robert, Y., Tchuente, M. Automata Networks in Computer Science: Theory and Applications. Manchester University Press, 1987.Google Scholar
  14. 14.
    Franklin, S., Garzon, M. Global dynamics in neural networks. Complex Systems 3 (1989), 29–36.zbMATHMathSciNetGoogle Scholar
  15. 15.
    Franklin, S., Garzon, M. Neural computability. In: Progress in Neural Networks 1 (ed. O. M. Omidvar). Ablex, Norwood, NJ, 1990. Pp. 128–144.Google Scholar
  16. 16.
    Funahashi, K.-I. On the approximate realization of continuous mappings by neural networks. Neural Networks 2 (1989), 183–192.CrossRefGoogle Scholar
  17. 17.
    Furst, M., Saxe, J. B., Sipser, M. Parity, circuits, and the polynomial-time hierarchy. Math. Systems Theory 17 (1984), 13–27.zbMATHMathSciNetCrossRefGoogle Scholar
  18. 18.
    Garzon, M., Franklin, S. Global dynamics in neural nets II. Report 89-9, Memphis State Univ., Dept. of Mathematical Sciences, 1989.Google Scholar
  19. 19.
    Garzon, M., Franklin, S. Neural computability II. In: Proc. of the 3rd Internat. Joint Conf. on Neural Networks, Vol. 1. IEEE, New York, 1989. Pp. 631–637.CrossRefGoogle Scholar
  20. 20.
    Godbeer, G. H., Lipscomb, J., Luby, M. On the Computational Complexity of Finding Stable State Vectors in Connectionist Models (Hopfield Nets). Technical Report 208/88, Dept. of Computer Science, Univ. of Toronto, March 1988.Google Scholar
  21. 21.
    Goldmann, M., Håstad, J., Razborov, A. Majority gates vs. general weighted threshold gates. In: Proc. of the 7th Ann. Conf. on Structure in Complexity Theory. IEEE, New York, 1992.Google Scholar
  22. 22.
    Goles, E., Fogelman, F., Pellegrin, D. Decreasing energy functions as a tool for studying threshold networks. Disrc. Appl. Math. 12 (1985), 261–277.CrossRefzbMATHGoogle Scholar
  23. 23.
    Goles, E., Martínez, S. Exponential transient classes of symmetric neural networks for synchronous and sequential updating. Complex Systems 3 (1989), 589–597.zbMATHMathSciNetGoogle Scholar
  24. 24.
    Goles, E., Martínez, S. Neural and Automata Networks. Kluwer Academic, Dordrecht, 1990.zbMATHGoogle Scholar
  25. 25.
    Goles, E., Olivos, J. The convergence of symmetric threshold automata. Info. and Control 51 (1981), 98–104.CrossRefzbMATHGoogle Scholar
  26. 26.
    Hajnal, A., Maass, W., Pudlák, P., Szegedy, M., Turán, G. Threshold circuits of bounded depth. In: Proc. of the 28th Ann. IEEE Symp. on Foundations of Computer Science. IEEE, New York, 1992. Pp. 99–110. (Revised version to appear in J. Comp. Syst. Sci.)Google Scholar
  27. 27.
    Haken, A. Connectionist networks that need exponential time to stabilize. Manuscript, 10 pp., January 1989.Google Scholar
  28. 28.
    Haken, A., Luby, M. Steepest descent can take exponential time for symmetric connection networks. Complex Systems 2 (1988), 191–196.zbMATHGoogle Scholar
  29. 29.
    Hartley, R., Szu, H. A comparison of the computational power of neural networks. In: Proc. of the 1987 Internat. Conf. on Neural Networks, Vol. 3. IEEE, New York, 1987. Pp. 15–22.Google Scholar
  30. 30.
    Hertz, J., Krogh, A., Palmer, R. G. Introduction to the Theory of Neural Computation. Addison-Wesley, Redwood City, CA, 1991.zbMATHGoogle Scholar
  31. 31.
    Hinton, G. E., Sejnowski, T. E. Learning and relearning in Boltzmann machines. In [62], pp. 282–317.Google Scholar
  32. 32.
    Hong, J. On connectionist models. Comm. Pure and Applied Math. XLI (1988), 1039–1050.Google Scholar
  33. 33.
    Hopfield, J. J. Neural networks and physical systems with emergent collective computational abilities. Proc. Nat. Acad. Sci. USA 79 (1982), 2554–2558.MathSciNetCrossRefGoogle Scholar
  34. 34.
    Hornik, K., Stinchcombe, M., White, H. Multilayer feedforward nets are universal approximators. Neural Networks 2 (1989), 359–366.CrossRefGoogle Scholar
  35. 35.
    Håstad, J. Almost optimal lower bounds for small depth circuits. In: Randomness and Computation. Advances in Computing Research 5 (ed. S. Micali). JAI Press, Greenwich, CT, 1989. Pp. 143–170.Google Scholar
  36. 36.
    Håstad, J., Goldmann, M. On the power of small-depth threshold circuits. Computational Complexity 1 (1991), 113–129.zbMATHMathSciNetCrossRefGoogle Scholar
  37. 37.
    Judd, J. S. On the complexity of loading shallow neural networks. J. Complexity 4 (1988), 177–192.zbMATHMathSciNetCrossRefGoogle Scholar
  38. 38.
    Judd, J. S. Neural Network Design and the Complexity of Learning. The MIT Press, Cambridge, MA, 1990.Google Scholar
  39. 39.
    Kamp, Y., Hasler, M. Recursive Neural Networks for Associative Memory. John Wiley & Sons, Chichester, 1990.zbMATHGoogle Scholar
  40. 40.
    Kearns, M., Valiant, L. G. Cryptographic limitations on learning Boolean formulae and finite automata. In: Proc. of the 21st Ann. ACM Symp. on Theory of Computing. ACM, New York, 1989. Pp. 433–444.Google Scholar
  41. 41.
    Kleene, S. C. Representation of events in nerve nets and finite automata. In: Automata Studies (ed. C. E. Shannon and J. McCarthy). Annals of Mathematics Studies n:o 34. Princeton Univ. Press, Princeton, NJ, 1956. Pp. 3–41.Google Scholar
  42. 42.
    Kohonen, T. Self-Organization and Associative Memory. 3rd Ed., Springer-Verlag, Berlin, 1989.zbMATHGoogle Scholar
  43. 43.
    Lin, J.-H., Vitter, J. S. Complexity results on learning by neural nets. Machine Learning 6 (1991), 211–230.Google Scholar
  44. 44.
    Maass, W., Schnitger, G., Sontag, E. D. On the computational power of sigmoid versus Boolean threshold circuits. In: Proc. of the 32nd Ann. IEEE Symp. on Foundations of Computer Science. IEEE, New York, 1991. Pp. 767–776.Google Scholar
  45. 45.
    McClelland, J. L., Rumelhart, D. E., et al. (eds.) Parallel Distributed Processing: Explorations in the Microstructure of Cognition, Vol. 2. The MIT Press, Cambridge, MA, 1986.Google Scholar
  46. 46.
    McCulloch, W. S., Pitts, W. A logical calculus of the ideas immanent in nervous activity. Bull. Math. Biophys. 5 (1943), 115–133. Reprinted in [4], pp. 18–27.zbMATHMathSciNetCrossRefGoogle Scholar
  47. 47.
    Minsky, M. L. Computation: Finite and Infinite Machines. Prentice-Hall, Englewood Cliffs, NJ, 1972.Google Scholar
  48. 48.
    Minsky, M. L., Papert, S. A. Perceptrons: An Introduction to Computational Geometry. The MIT Press, Cambridge, MA, 1969 (expanded edition 1988).zbMATHGoogle Scholar
  49. 49.
    Muroga, S. Threshold Logic and Its Applications. John Wiley & Sons, New York, 1971.zbMATHGoogle Scholar
  50. 50.
    Muroga, S., Toda, I., Takasu, S. Theory of majority decision elements. J. Franklin Inst. 271 (1961), 376–418.zbMATHMathSciNetCrossRefGoogle Scholar
  51. 51.
    Myhill, J., Kautz, W. H. On the size of weights required for linear-input switching functions. IRE Trans. Electronic Computers 10 (1961), 288–290.CrossRefGoogle Scholar
  52. 52.
    Obradovic, Z., Parberry, I. Analog neural networks of limited precision I: Computing with multilinear threshold functions (Preliminary version). In: Advances in Neural Information Processing Systems 2 (ed. D. S. Touretzky). Morgan Kaufmann, San Mateo, CA, 1990. Pp. 702–709.Google Scholar
  53. 53.
    Parberry, I. A primer on the complexity theory of neural networks. In: Formal Techniques in Artificial Intelligence: A Sourcebook (ed. R. B. Banerji). Elsevier-North-Holland, Amsterdam, 1990. Pp. 217–268.Google Scholar
  54. 54.
    Parberry, I., Schnitger, G. Relating Boltzmann machines to conventional models of computation. Neural Networks 2 (1989), 59–67.CrossRefGoogle Scholar
  55. 55.
    Pollack, J. On Connectionist Models of Natural Language Processing. Ph. D. Thesis, Univ. Illinois, Urbana, 1987.Google Scholar
  56. 56.
    Porat, S. Stability and looping in connectionist models with asymmetric weights. Biol. Cybern. 60 (1989), 335–344.zbMATHMathSciNetCrossRefGoogle Scholar
  57. 57.
    Raghavan, P. Learning in threshold networks. In: Proc. of the 1988 Workshop on Computational Learning Theory (ed. D. Haussler, L. Pitt). Morgan Kaufmann, San Mateo, CA, 1988. Pp. 19–27.Google Scholar
  58. 58.
    Reif, J. On threshold circuits and polynomial computation. In: Proc. of the 2nd Ann. Conf. on Structure in Complexity Theory. IEEE, New York, 1987. Pp. 118–123.Google Scholar
  59. 59.
    F. Rosenblatt. Priciples of Neurodynamics. Spartan Books, New York, 1962.Google Scholar
  60. 60.
    Roychowdhury, V., Siu, K. Y., Orlitsky, A., Kailath, T. A geometric approach to threshold circuit complexity. In: Proc. of the 4th Ann. Workshop on Computational Learning Theory (ed. L. G. Valiant, M. K. Warmuth). Morgan Kaufmann, San Mateo, CA, 1991. Pp. 97–111.Google Scholar
  61. 61.
    Rumelhart, D. E., Hinton, G. E., Williams, R. J. Learning internal representations by error propagation. In [62]. pp. 318–362.Google Scholar
  62. 62.
    Rumelhart, D. E., McClelland, J. L., et al. (eds.) Parallel Distributed Processing: Explorations in the Microstructure of Cognition, Vol. 1. The MIT Press, Cambridge, MA, 1986.Google Scholar
  63. 63.
    Siegelman, H. T., Sontag, E. D. On the computational power of neural nets. Report SYCON-91-11, Rutgers University, New Brunswick, NJ, Nov. 1991.Google Scholar
  64. 64.
    Siu, K.-Y., Bruck, J. Neural computation of arithmetic functions. Proc. of the IEEE 78 (1990), 1669–1675.CrossRefGoogle Scholar
  65. 65.
    Siu, K.-Y., Bruck, J. On the power of threshold circuits with small weights. SIAM J. Discr. Math. 4 (1991), 423–435.zbMATHMathSciNetCrossRefGoogle Scholar
  66. 66.
    Wegener, I. The Complexity of Boolean Functions. John Wiley & Sons, Chichester, and B. G. Teubner, Stuttgart, 1987.zbMATHGoogle Scholar
  67. 67.
    Yao, A. C. Separating the polynomial-time hierarchy by oracles. In: Proc. of the 26th Ann. IEEE Symp. on Foundations of Computer Science. IEEE, New York, 1985. Pp. 1–10.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1992

Authors and Affiliations

  • Pekka Orponen
    • 1
  1. 1.Department of Computer ScienceUniversity of HelsinkiHelsinkiFinland

Personalised recommendations