Distributed neural networks have been very succesful in modeling isolated cognitive phenomena, but complex high-level behavior has been tractable only with symbolic artificial intelligence techniques. Aiming to bridge this gap, this book describes DISCERN, a complete natural language processing system implemented entirely at the subsymbolic level. In DISCERN, distributed neural network models of parsing, generating, reasoning, lexical processing and episodic memory are integrated into a single system that learns to read, paraphrase, and answer questions about stereotypical narratives. The book introduces a general approach to building high-level cognitive models from distributed neural networks, and shows how the special properties of such networks provide an advantage in modeling human performance. In this approach, connectionist networks are not only plausible models of isolated cognitive phenomena, but also sufficient constituents for complete artificial intelligence systems. The book, which includes a comprehensive survey of the connectionist literature related to natural language processing, is intended to researchers interested in practical techniques for high-level representation, inferencing, memory modeling, and modular connectionist architectures. The source code and data for DISCERN are available by ftp, and can be used as a starting point for further experiments with connectionist high-level AI.
[ The full text of the book is not available electronically. However, you can access the precis that appeared in the electronic journal Psycoloquy 94.5.46 (1994), followed by multiple reviews and responses, and the dissertation on which the book expands. ]
Cambridge, MA, 1993. MIT Press.