A Connectionist Realization Applying Knowledge-Compilation and Auto-Segmentation in a Symbolic Assignment Problem
A symbolic assignment problem has been solved by making use of the fact that it can be represented as a decomposable production system. We explain how the structure of the given constraint satisfaction problem (CSP) can be exploited to design a two component architecture of a neural net interacting with a scheduler. We describe the relation of problem parameters to the net design, using a feedforward net with error backpropagation. Different versions of the net design are contrasted. We discuss the advantages of our architecture and relate the results of the connectionist approach to a solving of the problem with backtrack search. The CSP was part of a case study, a knowledge-based system for the automatic configuration of telephone exchanges. An enlargement of the architecture and its application is foreseen.
Unable to display preview. Download preview PDF.
- Barnden, John A., Neural-Net Implementation of Complex Symbol-Processing in a Mental Model Approach to Syllogistic Reasoning, Proceedings of IJCAI-89, Detroit, Mich., 1989, 568–573.Google Scholar
- Ekeberg, O., and A. Lansner, Automatic Generation of Internal Representations in a Probabilistic Artificial Neural Network, in: Neural Networks from Models to Applications ( Personnaz, L., and G. Dreyfus, eds.), Paris: IDSET, 1989, 179–186.Google Scholar
- Kaindl, H., and H.G. Ziegeler, Some Aspects of Knowledge-Based Configuration, Proceedings AVIGNON ‘80: Expert systems & their applications, Specialized Conference: Artificial Intelligence, Telecommunications & Computer Systems, 1990, 41–54.Google Scholar
- Mackworth, A.K., Constraint Satisfaction, in Encyclopedia of Artificial Intelligence ( Shapiro, S.C., ed.), New York, N.Y.: Wiley, 1987, 205–211.Google Scholar
- Quinlan, J.R., Learning efficient classification procedures and their application to chess end games: in: Machine Learning 2 (Michalski, R.S., J.G. Carbonell, and T.M. Mitchell, eds.), Palo Alto, Ca.: Tioga, 1984, 463–482.Google Scholar
- Rumelhart, D., J. McClelland et al. (ed.), Parallel Distributed Processing: Explorations in the Microstructure of Cognition, Cambridge, Ma.: MIT Press, 1986.Google Scholar
- Tagliarini, G.A. and E.W. Page, Solving Constraint Satisfaction Problems with Neural Networks, Proc. IEEE First International Conference on Neural Networks, San Diego, 1987.Google Scholar
- Ziegeler, H.G., and H. Kaindl, A Cyclic Pattern Resulting from a Constraint Satisfaction Search, working paper, to be presented at the AAAI-90 Workshop on Constraint Directed Reasoning, July 1990.Google Scholar