Connectionism in Context aims to broaden and extend the debate concerning the significance of connectionist models. The volume collects together a variety of perspectives by experimental and developmental psychologists, philosophers and active AI researchers. These contributions relate con- nectionist ideas to historical psychlogical debates, e.g., over behaviourism and associationism, to develop- mental and philosophical issues. The result is a volume which addresses both familiar, but central, topics such as the relation between connectionism and classical AI, and less familiar, but highly challenging topics, such as connectionism,associationism and behaviourism, the dis- tinction between perception and cognition, the role of en- vironmental structure, and the potential value ofconnec- tionism as a means of "symbol grounding". The nine essays have been written with an interdisciplinary audience in mind and avoid both technical jargon and heavy mathematics.
Connectionism is currently one of the most flourishing and interdisciplinary areas of cognitive science. Drawing on research in neural computation and networks it has found applications in areas such as psychology and animal intelligence. By using types of network which attempt to mirror our own cognitive architecture, connectionism is making breakthroughs in the understanding of the human mind a real possibility.
Part of a series on cognitive behaviour and science, based on a 1990 conference sponsored by the Cognitive Science Program and the Linguistics Department of Simon Fraser University, Vancouver, British Columbia, Canada.
Presenting research on the computational abilities of connectionist, neural, and neurally inspired systems, this series emphasizes the question of how connectionist or neural network models can be made to perform rapid, short-term types of computation that are useful in higher level cognitive processes. The most recent volumes are directed mainly at researchers in connectionism, analogy, metaphor, and case-based reasoning, but are also suitable for graduate courses in those areas.
This series will include monographs and collections of studies devoted to the investigation and exploration of knowledge, information and data processing systems of all kinds, no matter whether human, (other) animal, or machine. Its scope is intended to span the full range of interests from classical problems in the philosophy of mind and philosophical psychology through issues in cognitive psychology and sociobiology (concerning the mental capabilities of other species) to ideas related to artificial intelligence and to computer science. While primary emphasis will be placed upon theoretical, conceptual and epistemological aspects of these problems and domains, empirical, experimental and methodological studies will also appear from time to time. One of the most, if not the most, exciting developments within cognitive science has been the emergence of connectionism as an alternative to the computational conception of the mind that tends to dominate the discipline. In this volume, John Tienson and Terence Horgan have brought together a fine collection of stimulating studies on connectionism and its significance. As the Introduction explains, the most pressing questions concern whether or not connectionism can provide a new conception of the nature of mentality. By focusing on the similarities and differences between connectionism and other approaches to cognitive science, the chapters of this book supply valuable resources that advance our understanding of these difficult issues. J.H.F.
In order to build "intelligent" machines, many researchers have turned to the only naturally occurring intelligent system: the brain. For quite a while now, both the function and architecture of the brain have served as inspiration to philosophers, psychologists, computer scientists, neurobiologists, physicists and others in their quest for solving problems that seem to require intelligence in their own particular domain. The progress in the field of connectionism -- or artificial neural networks -- has had its ups and downs during its maturing years. Advocates of the field pointed out the virtues of connectionist systems, dealing with low-level cognitive tasks such as visual recognition and pattern completion, and inherent properties such as generalization, fault tolerance and parallel processing. However, research in the field virtually came to a halt at the end of the 1960s when Minsky and Papert published their critical analysis of connectionist systems, Perceptrons. In the beginning of the 1980s, the field was reborn with the appearance of new powerful learning methods which overcame many of the computational problems identified by Minsky and Papert. This volume is characterized by a number of different research directions distinguished by their perspectives on systems comprising interconnected sets of simple processing elements. Scientists who have strong backgrounds in neurobiology concentrate on the issues involved when modelling natural systems. Researchers with philosophical and psychological backgrounds stress other aspects which might not always be intuitively relevant to biology but instead are concerned with the mind and its higher-order cognitive capabilities. On the other hand, many researchers and engineers in industry take advantage of the wide applicability and mathematical properties of connectionist systems in order to solve practical problems, sacrificing even more of the principles underlying the basic idea of mimicking the function and architecture of the brain. None of these directions are right or wrong, but there has perhaps been too little exchange of knowledge and experience between them. The main purpose for organizing this conference was to bring together researchers with different backgrounds to exchange ideas and visions in the broad field of connectionism -- providing means for new insights that may push this area to another major breakthrough.
A variety of ideas, approaches, and techniques exist -- in terms of both architecture and learning -- and this abundance seems to lead to many exciting possibilities in terms of theoretical advances and application potentials. Despite the apparent diversity, there is clearly an underlying unifying theme: architectures that bring together symbolic and connectionist models to achieve a synthesis and synergy of the two different paradigms, and the learning and knowledge acquisition methods for developing such architectures. More effort needs to be extended to exploit the possibilities and opportunities in this area. This book is the outgrowth of The IJCAI Workshop on Connectionist-Symbolic Integration: From Unified to Hybrid Approaches, held in conjunction with the fourteenth International Joint Conference on Artificial Intelligence (IJCAI '95). Featuring various presentations and discussions, this two-day workshop brought to light many new ideas, controversies, and syntheses which lead to the present volume. This book is concerned with the development, analysis, and application of hybrid connectionist-symbolic models in artificial intelligence and cognitive science. Drawing contributions from a large international group of experts, it describes and compares a variety of models in this area. The types of models discussed cover a wide range of the evolving spectrum of hybrid models, thus serving as a well-balanced progress report on the state of the art. As such, this volume provides an information clearinghouse for various proposed approaches and models that share the common belief that connectionist and symbolic models can be usefully combined and integrated, and such integration may lead to significant advances in understanding intelligence.
The rapid growth of neural network research has led to a major reappraisal of many fundamental assumptions in cognitive and perceptual psychology. This text—aimed at the advanced undergraduate and beginning postgraduate student—is an in-depth guide to those aspects of neural network research that are of direct relevance to human information processing. Examples of new connectionist models of learning, vision, language and thought are described in detail. Both neurological and psychological considerations are used in assessing its theoretical contributions. The status of the basic predicates like exclusive-OR is examined, the limitations of perceptrons are explained and properties of multi-layer networks are described in terms of many examples of psychological processes. The history of neural networks is discussed from a psychological perspective which examines why certain issues have become important. The book ends with a general critique of the new connectionist approach. It is clear that new connectionism work provides a distinctive framework for thinking about central questions in cognition and perception. This new textbook provides a clear and useful introduction to its theories and applications.
The latest title in the Cognitive Science and Second Language Acquisition Series presents a comprehensive review of connectionist research in second language acquisition (SLA). Second language researchers and the cognitive science community will find accessible discussions of the relevance of connectionist research to SLA. This important volume is key reading for any student or researcher interested in how second language acquisition can be better understood from a connectionist perspective.
