No CrossRef data available.
Article contents
Learning Simple Things: A Connectionist Learning Problem from Various Perspectives
Published online by Cambridge University Press: 28 February 2022
Extract
The performance of a connectionist learning system on a simple problem has been described by Hinton and is briefly reviewed here: a finite function is learned, and the system generalizes correctly from partial information by finding simple “features” of the environment. For comparison, a very similar problem is formulated in the Gold paradigm of discrete learning functions. Identification in the limit from positive text of a large class of functions including Hinton's is achievable with a trivial, conservative learning strategy. Using Valiant's approach, we place an arbitrary finite bound on function complexity and then we can guarantee text and resource efficiency relative to a probabilistic criterion of success. But the connectionist system generalizes. That is, it uses a non-conservative learning strategy. We define a simple, non-conservative strategy that also generalizes like the connectionist system, finding simple “features” of the environment.
- Type
- Part XIII. Formal Learning Theory
- Information
- Copyright
- Copyright © 1989 by the Philosophy of Science Association
Footnotes
1
I am grateful to Geoffrey Hinton, Edward Stabler, William Demopoulos, Daniel Osherson, and Zenon Pylyshyn for helpful discussions of this material. This research was supported in part by the Canadian Institute for Advanced Research and the Natural Science and Engineering Research Council of Canada.
References
Brayton, R.K. and McMullin, C. (1982), “The Decomposition and Factorization of Boolean Expressions.” Procs. 1982 IEEE Int. Symp. on Circuits and Systems: 49-54.Google Scholar
de Micheli, G., Brayton, R.K. and Sangiovanni-Vincentelli, A. (1985), “Optimal State Assignment for Finite State Machines.” IEEE Transactions on Computer-Aided Design, CAD-4(3): 269–285.CrossRefGoogle Scholar
Gold, E.M. (1967), “Language Identification in the Limit.” Information and Control 10: 447–474.CrossRefGoogle Scholar
Goodman, N. (1955), Fact, Fiction and Forecast. Cambridge, Massachusetts: Harvard University Press.Google Scholar
Hartmanis, J. and Stearns, R.E. (1966), Algebraic Structure Theory of Sequential Machines. Englewood Cliffs, NJ: Prentice-Hall.Google Scholar
Hinton, G.E. (1986), “Learning Distributed Representations of Concepts.” Procs. of the 8th Ann. Conf. of the Cognitive Science Society: 1-12. New Jersey: Erlbaum.Google Scholar
Hinton, G.E. (1987ms), “Connectionist Learning Procedures.” Forthcoming in Artificial Intelligence.Google Scholar
Johnson, R.A. and Wichern, D.W. (1988), Applied Multivariate Statistical Analysis, 2nd Edition. Englewood Cliffs: Prentice-Hall.Google Scholar
Osherson, D.N. (1978), “Three Conditions on Conceptual Naturalness.” Cognition 6: 263–289.CrossRefGoogle Scholar
Osherson, D.N., Stob, M. and Weinstein, S. (1986), Systems that Learn: An Introduction to Learning Theory for Cognitive and Computer Scientists. Cambridge, Massachusetts: MIT Press.Google Scholar
Pinker, S. and Prince, A. (1988), “On Language and Connectionism: Analysis of a Distributed Model of Language Acquisition.” Cognition 28: 73–193.CrossRefGoogle ScholarPubMed
Pylyshyn, Z.W. (1988ms) “The Role of Cognitive Architecture in Theories of Cognition.” Forthcoming in VanLehn, K., ed., Architectures for Intelligence.Google Scholar
Quine, W.V. (1952), “The Problem of Simplifying Truth Functions.” American Mathematical Monthly, 59: 521–531.CrossRefGoogle Scholar
Stabler, E.P. Jr. (1989), “Learning Simple Things: A Connectionist Learning Problem from Various Perspectives.” Technical Report, Department of Computer Science, University of Western Ontario.Google Scholar
Valiant, L.G. (1984), “A Theory of the Learnable.” Communications of the ACM 27:1134–1142.CrossRefGoogle Scholar