flach@kub.nl (Peter Flach) (04/22/91)
Book Announcement FUTURE DIRECTIONS IN ARTIFICIAL INTELLIGENCE IFIP TC12 Founding Workshop Collected Papers Edited by Peter A. Flach and Robert A. Meersman Tilburg University, the Netherlands Published by North-Holland (Amsterdam), February 1991 xiv+190 pages ISBN 0-444-89048-3 This book has its roots in an informal workshop enti- tled 'The Future of Research in Artificial Intelligence', which was held at the occasion of the Eleventh International Joint Conference on Artificial Intelligence, August 1989, Detroit. The workshop brought together a number of interna- tional AI specialists, with the goal of discussing important future research directions in Artificial Intelligence. More- over, the workshop resulted in the establishment of Techni- cal Committee 12 on Artificial Intelligence, of the Interna- tional Federation for Information Processing (IFIP). The book contains post-edited versions of the talks presented at the workshop, along with a number of solicited papers. The 16 papers from 21 authors in this volume con- tain a number of original, thought-provoking, sometimes controversial viewpoints. The book is divided into four main parts: Methodology, Paradigms, Trends, and Prospects. The first part, Methodology, contains three papers, discuss- ing the following topics: how should AI students be edu- cated; AI's place in relation to Computer Science; and the increasing specialisation of AI research. The next part con- sists of four papers discussing the paradigms that dominated AI research in the past, such as the symbol processing para- digm, and a perceived need for new paradigms. The third part contains four papers outlining the state of the art and trends in knowledge representation, intelligent interfaces, intelligent networking systems, and distributed AI. The fourth part, Prospects, contains four papers which look somewhat further in the future, suggesting important research topics, and discussing the future of AI research in general. A list of references and an extensive index con- clude the book. ============================================================ TABLE OF CONTENTS List of Contributors Editors' preface Part I --METHODOLOGY The need for a formal education in Artificial Intelligence Ranan B. Banerji 1. Experimental Computer Science? 2. The early days 3. Some symptoms 4. The way it hurts 5. A plea to the schools Informatics as the scientific environment for Artificial Intelligence Jozef Gruska 1. Introduction 2. AI and Computer Science -- their ups and downs 3. Informatics as a new fundamental science and new methodology for science and technology 4. Implications for AI 5. Conclusions Towards a reintegration of Artificial Intelligence research John F. Sowa 1. Effects of specialization 2. Towards a reintegration Part II -- PARADIGMS Paradigm shifts in Artificial Intelligence Armand de Callatay 1. The changing environment for AI 1.1 What was AI? 1.2 AI and Computer Science 2. Paradigm shifts 2.1 New definition of a 'simple' computing instruction 2.2 Processing directly in memory 2.3 Computation on processors which are not universal 2.4 Limitations of rule-based systems 2.5 Reviving 'system theory' 2.6 Primitives and emergent properties 2.7 Knowledge base of procedures 2.8 Study of defective machines 2.9 Learning by experience as in animals 3. Research related to brains 4. New computer architectures for AI 5. AI research on hybrid systems 6. A future direction for AI research 7. Conclusions. Why today's computers don't learn the way people do William J. Clancey 1. Introduction 2. Learning in AI programs 3. An alternative view 4. Two examples 5. Summary Future directions of Artificial Intelligence in a resource-limited environment Ryszard S. Michalski & David C. Littman 1. Introduction 2. An outline of future intelligent systems 2.1 Functionality 2.2 Inference capabilities 2.3 Engineering considerations 3. Paradigms for AI research 4. Symbolic preeminence 5. Conclusions Beyond the symbolic paradigm Leon S. Sterling, Randall D. Beer & Hillel J. Chiel 1. Introduction 2. Coping with a changing, unpredictable environment 3. Intelligence as adaptive behavior 4. Lessons from biology 5. Computational ethology 6. Conclusions Part III -- TRENDS Research trends in knowledge representation Luigia Carlucci Aiello & Daniele Nardi 1. Premise 2. The problem 3. Hybrid reasoning 4. Reasoning with incomplete knowledge 5. Reasoning with contradictory knowledge 6. Reasoning about knowledge and reasoning 'Artificial' interfaces for knowledge acquisition: a futuristic scenario R. Chandrasekar & S. Ramani 1. Introduction 2. Limitations in current AI systems 3. Prescription for the future 4. The structure of the ideal interface 4.1 New modes of control and communication 4.2 Intelligent interfaces 4.3 The case for multi-modal communication 4.4 New devices: a peek at what is possible 5. Lessons from child language research 5.1 Language behaviour 5.2 What Motherese offers: an aid to learning in children 5.3 Motherese: where else can it be applied? 5.4 The cognitive development of computer programs 6. An example: a natural language understanding system 7. Conclusions Intelligent distributed and networking systems Jacek Maitan 1. Current status 1.1 Object-oriented model of distributed systems 1.2 Implementation problems 2. Emerging needs 3. Scope of the problem 4. Research directions 4.1 Nature of distributed problems 4.2 Implementation and maintenance of distributed systems 4.3 Theoretical models of computation and communication 5. Summary Distributed Artificial Intelligence Zhongzhi Shi 1. Introduction 2. The key issues 2.1 Parallel distributed processing 2.2 Knowledge representations 2.3 Task decomposition 2.4 Cooperative strategies 3. The principle techniques 3.1 Coordinating via organizational structuring 3.2 Contract network 3.3 Task centralization 3.4 Partial global planning 3.5 Distributed knowledge base management system 4. Applications 5. Conclusions Part IV -- PROSPECTS On the future (and present) state of Artificial Intelligence Yves Kodratoff 1. Existing features that will become better acknowledged 1.1 Knowledge intensiveness 1.2 Transparent box and explanations 2. Existing topics that will develop 2.1 Knowledge intensive deduction 2.2 Knowledge intensive induction 3. Future topics 3.1 Explanations 3.2 Multi-agent interactions 3.3 Symbolic vs. numeric 3.4 Analogy 4. Applications 4.1 Application of AI to vision 4.2 Application of expert systems by naive users 4.3 Certification of expert systems by machine learning techniques 5. Conclusions Artificial Intelligence needs its Eisenstein and Chaplin Andras Markus & Elod Knuth 1. The personification of software 1.1 Individual features 1.2 The programUs lost identity and the need for archivation 2. The growing importance of transformations between representations 3. New integration and separation of cognitive activities 3.1 Generating natural phenomena 3.2 AI for and against the deterioration of traditional skills AI technology, non-existent or extinct? Peter van Lith The future of research in Artificial Intelligence Laurent Siklossy 1. Introduction 2. The future of research 3. The future of AI 4. Research in AI 5. Future research in AI 6. Conclusions: more of the same! EPILOGUE The dialectics of Artificial Intelligence Peter A. Flach 1. Introduction 2. What is Artificial Intelligence? 3. Is the mind a computer program? 4. Are sub-symbolic representations necessary? 4.1 Scientific paradigms 4.2 Symbolism vs. connectionism 5. Conclusions References Index -- Peter A. Flach Institute for Language Technology INTERNET: flach@kub.nl and Artificial Intelligence (ITK) BITNET: flach@htikub5 Tilburg University, PObox 90153 (+31) (13) 66 3119 5000 LE Tilburg, the Netherlands