Representations of Commonsense Knowledge (Morgan Kaufmann Series in Representation and Reasoning)

A central goal of artificial intelligence is to give a computer program commonsense understanding of basic domains such as time, space, simple laws of nature, and simple facts about human minds. Many different systems of representation and inference have been developed for expressing such knowledge and reasoning with it. Representations of Commonsense Knowledge is the first thorough study of these techniques.The first three chapters establish a general framework in domain-independent terms, discussing methodology, deductive logics, and theories of plausible inference. Subsequent chapters each deal with representation and inferences in specific domains: quantities, time, space, physics, knowledge and belief, plans and goals, and interactions among agents. The power of these representations in expressing world knowledge and in supporting significant inferences is analyzed using many detailed examples. The discussion includes both representations that have been used in successful AI programs and those that have been developed in purely abstract settings.Representations of Commonsense Knowledge is an essential reference for AI researchers and developers. It can also be used as a textbook in advanced undergraduate or graduate courses. Each chapter contains exercises and suggestions for further reading. Readers who have completed it will be prepared to read original technical papers in the area and to begin their own work in developing useful representations for AI programs. A central goal of artificial intelligence is to give a computer program commonsense understanding of basic domains such as time, space, simple laws of nature, and simple facts about human minds. Many different systems of representation and inference have been developed for expressing such knowledge and reasoning with it. "Representations of Commonsense Knowledge" is the first thorough study of these techniques. The first three chapters of the book establish a general framework in domain-independent terms, discussing methodology, deductive logics, and theories of plausible inference. Subsequent chapters each deal with representations and inferences in specific quantities, time, space, physics, knowledge and belief, plans and goals, and interactions among agents. The power of these representations in expressing world knowledge and in supporting significant inferences is analyzed using many detailed examples. The discussion includes both representations that have been used in successful AI programs and those that have been developed in purely abstract settings. Representations of Commonsense Knowledge is an essential reference for AI researchers and developers. It can also be used as a textbook in advanced undergraduate or graduate courses; each chapter contains exercises and suggestions for further reading. Readers who have completed the book will be prepared to read original technical papers in the area and to begin their own work in developing useful representations for AI programs. "A central goal of artificial intelligence is to give a computer program commonsense understanding of basic domains such as time, space, simple laws of nature, and simple facts about human minds. Many different systems of representation and inference have been developed for expressing such knowledge and reasoning with it. Representations of Commonsense Knowledge is the first thorough study of these techniques. The first three chapters establish a general framework in domain-independent terms, discussing methodology, deductive logics, and theories of plausible inference. Subsequent chapters each deal with representations and inferences in specific domains: quantities, time, space, physics, knowledge and belief, plans and goals, and interactions among agents. The power of these representations in expressing world knowledge and in supporting significant inferences is analyzed using many detailed examples. The discussion includes both representations that have been used in successful AI programs and those that have been developed in purely abstract settings."--Book cover Representations of Commonsense Knowledge provides a rich language for expressing commonsense knowledge and inference techniques for carrying out commonsense knowledge. This book provides a survey of the research on commonsense knowledge. Organized into 10 chapters, this book begins with an overview of the basic ideas on artificial intelligence commonsense reasoning. This text then examines the structure of logic, which is roughly analogous to that of a programming language. Other chapters describe how rules of universal validity can be applied to facts known with absolute certainty to deduce other facts known with absolute certainty. This book discusses as well some prominent issues in plausible inference. The final chapter deals with commonsense knowledge about the interrelations and interactions among agents and discusses some issues in human and social interactions that have been studied in the artificial intelligence literature. This book is a valuable resource for students on a graduate course on knowledge representation. Front Cover; Representations of Commonsense Knowledge; Copyright Page; Dedication; Preface; Table of Contents; List of Tables; List of Named Axioms; Chapter 1. Automating Common Sense; 1.1 Knowledge Bases; 1.2 Methodology; 1.3 Implementation; 1.4 The Role of Natural Language; 1.5 The Role of Logic; 1.6 Incomplete and Uncertain Knowledge; 1.7 Vagueness; 1.8 Indexicals; 1.9 Commonsense Reasoning in Artificial Intelligence; 1.10 Philosophy; l.11 Mathematics and Commonsense Reasoning; 1.12 References; Chapter 2. Logic; 2.1 Logical Systems and Languages; 2.2 Propositional Calculus The student papers selected for this volume have been revised to incorporate feedback the students received from their peers during the program, and afterwards from their assigned editor (one of us). A few papers describe work that actually took place during the summer school, or immediately after it. The faculty papers are likewise timely and of very high quality. Readers wishing to monitor the neural network scene will find there is much to learn about the current state of the art The summer school program was held at Carnegie Mellon U., June 1988. The pape (updated and revised for publication) present a broad yet detailed view of current research in optimization methods, learning theory, knowledge representation, vision, speech, cognitive science, architectures, and hardwa