Abstract
We report on an investigation of the feasibility of using connectionist-nets of the Rumelhart semi-linear feedforward type for learning and exercising intelligent system control. Of particular interest is the ability of such nets to discover the basis for decision and control. Of great interest also is an unexpected finding that such nets automatically build up a great deal of redundancy within itself so that it has a certain holographic quality to it and its performance degrades gradually with internal damage.
Similar content being viewed by others
References
Burr, D. J. (1986) A neural network digit recognizer, Proc. 1986 IEEE Conference on Systems, Man & Cybernetics, Atlanta, GA., 14–17 Oct. 1986.
Hopfield, J. J. (1982) Neural networks and physical systems with emergent collective computational abilities, Proc. Natl. Acad. Sci. USA, Biophysics 79, 2554–2558.
Hopfield, J. J. (1984) Neurons with graded response have collective computational properties like those of two-state neurons, Proc. Natl. Acad. Sci. USA, Biophysics 81, 3088–3092.
Hopfield, J. J. and Tank, D. W. (1985) ‘Neural’ computation of decisions in optimization problems, Biological Cybernetics 52, 141–152.
Hu, Chi-Heng and Pao, Yoh-Han, (1985) Processing of pattern-based information, Part II: Descriptions of inductive inference in terms of transition networks, Advances in Information Systems Science, Vol. 9 (ed. Julius T. Tou), Plenum Press, Chap. 4, pp. 261–286.
Michalski, Ryszard, S. (1980) Pattern recognition as rule-guided inference, IEEE Trans. on Pattern Analysis and Machine Intelligence, Vol. 2, No. 4.
Pao, Yoh-Han and Hu, Chi-Heng, (1985) Processing of pattern-based information, Part I: inductive inference methods suitable for use in pattern recognition and artificial intelligence, Advances in Information Systems Science, Vol. 9 (ed. Julius T. Tou), Plenum Press, Chap. 3, pp. 221–260 (1985).
Pao, Yoh-Han and Sobajic, D. J. (1987a) A perspective on the role of parallel distributed processing of the neural-net type in implementing automation, 1987 IEEE Workshop on Languages for Automation, Vienna, Austria, 24–27 August 1987.
Pao, Yoh-Han and Sobajic, D. J. (1987b) Connectionist-net technology for intelligent robotics control, International Workshop on Robotics: Trends, Technology and Applications, Madrid, Spain, 30 September–1 October 1987.
Quinlan, J. R. (1983) Learning efficient classification procedures and their application to chess end games Machine Learning: An Artificial Intelligence Approach (eds. R. S. Michalski, J. G. Carbonell, and T. M. Mitchell), Tioga, Palo Alto.
Rumelhart, D. E., Hinton, G. E. and Williams, R. J. (1986) Learning internal representations by error propagation in parallel distributed processing, Explorations in the Microstructures of Cognition, Vol. 1: Foundations, MIT Press, pp. 318–362.
Winston, Patrick Henry, (1975) Learning structured descriptions from examples, The Psychology of Computer Vision, (ed. Patrick H.Winston), McGraw-Hill, New York (1975).
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Pao, YH. Autonomous machine learning of effective control strategies with connectionist-net. J Intell Robot Syst 1, 35–53 (1988). https://doi.org/10.1007/BF00437319
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF00437319