Parallel distributed processing: explorations in the microstructure of cognition, vol. 1
Connectionist inference models
Neural Networks
A connectionist model for predicate logic reasoning using coarse-coded distributed representations
KES'05 Proceedings of the 9th international conference on Knowledge-Based Intelligent Information and Engineering Systems - Volume Part II
| Hi-index | 0.00 |
Traditionally reasoning systems have been implemented using symbolic methods of artificial intelligence. Connectionist methods of implementing reasoning systems form an alternative paradigm. Among the connectionist reasoning systems two types of representational methods can be used. They are i) localist and ii) distributed representational methods. In the literature, some localist methods for reasoning were used in connectionist systems. Since those systems used localist representations, advantages of distributed representations are not obtainable by them. In this paper, we describe the design and implementation of a connectionist knowledge based system which integrates a connectionist predicate logic reasoning system and a connectionist semantic network. The system uses distributed coarse-coded representations. The connectionist predicate logic system supports both simple rules as well as a complex rule having multiple conjunctions. Distributed representations have advantages of increased fault tolerance, graceful degradation of performance; neural plausibility, cognitive modeling and parallel distributed processing. The system besides showing above features allows the communication between these two connectionist systems and makes it possible to access the information of attributes and corresponding values from the connectionist semantic network for the entities used in the connectionist predicate logic system.