ABSTRACT
Connectionist models allow for much easier integration of modules than is possible with symbolic/heuristic search-based systems. Symbolic systems require either a very simple architecture or a sophisticated communications facility in order to build a system composed of many modules. Learning is one of the most exciting aspects of connectionist models for both the Artificial Intelligence and psychology communities. Connectionist systems for language processing have assumed that sentences will be preceded and followed by quiescent periods. Connectionist models served as initial inspiration to designers of new generation hardware, though many parallel architectural ideas were already being explored in the pursuit of greater speed. Natural language processing work has suffered and still suffers from a shortage of good ideas for applications. One of the fundamental problems of natural language processing is word sense disambiguation. Determining the correct sense of a word for a particular use involves the interaction of many sources of knowledge: syntactic, semantic, and pragmatic.
