Mueller Commonsense Reasoning

An Event Calculus Based Approach
2. Auflage 2014
ISBN: 978-0-12-801647-3
Verlag: William Andrew Publishing
Format: EPUB
Kopierschutz: 6 - ePub Watermark

E-Book, Englisch, 516 Seiten, Format (B × H): 191 mm x 235 mm

To endow computers with common sense is one of the major long-term goals of artificial intelligence research. One approach to this problem is to formalize commonsense reasoning using mathematical logic. Commonsense Reasoning: An Event Calculus Based Approach is a detailed, high-level reference on logic-based commonsense reasoning. It uses the event calculus, a highly powerful and usable tool for commonsense reasoning, which Erik Mueller demonstrates as the most effective tool for the broadest range of applications. He provides an up-to-date work promoting the use of the event calculus for commonsense reasoning, and bringing into one place information scattered across many books and papers. Mueller shares the knowledge gained in using the event calculus and extends the literature with detailed event calculus solutions that span many areas of the commonsense world.

The Second Edition features new chapters on commonsense reasoning using unstructured information including the Watson system, commonsense reasoning using answer set programming, and techniques for acquisition of commonsense knowledge including crowdsourcing.



Drawing upon years of practical experience and using numerous examples and illustrative applications Erik Mueller shows you the keys to mastering commonsense reasoning. You'll be able to:

- Understand techniques for automated commonsense reasoning
- Incorporate commonsense reasoning into software solutions
- Acquire a broad understanding of the field of commonsense reasoning.
- Gain comprehensive knowledge of the human capacity for commonsense reasoning

Zielgruppe

<p>Researchers and graduate students in the areas of cognitive science and artificial intelligence</p>

Autoren/Hrsg.

Mueller, Erik T.

Fachgebiete

• Mathematik | Informatik EDV | Informatik Informatik Künstliche Intelligenz
• Interdisziplinäres Wissenschaften Wissenschaften Interdisziplinär Neurowissenschaften, Kognitionswissenschaft

Inhaltsverzeichnis

Part I: Foundations

Part II: Commonsense Phenomena

Part III: Commonsense Domains

Part IV: Default Reasoning

Part V: Programs and Applications

Part VI: Logical and Non-logical Methods

Part VII: Knowledge Acquisition

Part VIII: Conclusion

Part IX: Appendices

Leseproben

Preface
Commonsense reasoning is the sort of reasoning we all perform about the everyday world. We can predict that, if a person enters a kitchen, then afterward the person will be in the kitchen. Or that, if someone who is holding a newspaper walks into a kitchen, then the newspaper will be in the kitchen. Because we make inferences such as these so easily, we might get the impression that commonsense reasoning is a simple thing. But it is very complex. Reasoning about the world requires a large amount of knowledge about the world and the ability to use that knowledge. We know that a person cannot be in two places at once, that a person can move from one location to another by walking, and that an object moves along with a person holding it. We have knowledge about objects, events, space, time, and mental states and can use that knowledge to make predictions, explain what we observe, and plan what to do. This book addresses the following question: How do we automate commonsense reasoning? In the last few decades, much progress has been made on this question by artificial intelligence researchers. This book provides a detailed account of this progress and a guide to automating commonsense reasoning using logic. We concentrate on one formalism, the event calculus, that incorporates many of the discoveries of the field. Although the event calculus is defined by a handful of classical predicate logic axioms, it enables reasoning about a wide range of commonsense phenomena. Why commonsense reasoning?
Why study commonsense reasoning? The first reason for studying commonsense reasoning is practical. Automated commonsense reasoning has many applications ranging from intelligent user interfaces and natural language processing to robotics and vision. Commonsense reasoning can be used to make computers more human-aware, easier to use, and more flexible. The second reason for studying commonsense reasoning is scientific. Common-sense reasoning is a core capability of intelligence that supports many other high-level capabilities. The ability to understand what is happening in a story, for example, crucially involves commonsense reasoning. By studying commonsense reasoning we can gain a greater understanding of what intelligence is. Approach
The approach to commonsense reasoning taken in this book is not shared by all researchers. My approach can be characterized by the following assumptions. I assume, along with most cognitive scientists, that commonsense reasoning involves the use of representations and computational processes that operate on those representations. I assume, along with researchers in symbolic artificial intelligence, that these representations are symbolic. I assume, along with researchers in logic-based artificial intelligence, that commonsense knowledge is best represented declaratively rather than procedurally. I use the declarative language of classical many-sorted predicate logic with equality. I do not claim that the methods for commonsense reasoning presented in this book are the methods used by humans. This book presents one way of automating commonsense reasoning. How humans perform commonsense reasoning is an interesting topic, but it is not the topic of this book. There is evidence both for and against the use of logic in human reasoning. Intended audience
This book is intended for use by researchers and students in the areas of computer science, artificial intelligence, mathematics, and philosophy. It is also intended for use by software designers wishing to incorporate commonsense reasoning into their applications. The book can be used as a graduate-level textbook for courses on commonsense reasoning and reasoning about action and change, as well as a reference work for researchers working in these areas. It will be of interest to those using logic as their primary technique, as well as those using other techniques. This book can also be used as a supplementary graduate-level or advanced undergraduate textbook for courses on knowledge representation and artificial intelligence. I assume the reader has some familiarity with predicate logic, although reviews of predicate logic are provided in Chapter 2 and Appendix A. Roadmap
This book consists of 20 chapters and six appendices. The chapters are organized into eight parts. Part I describes the foundations of the event calculus. Part II deals with various commonsense phenomena. Chapter 3 discusses the effects of events. Chapter 4 discusses the triggering of events by conditions. Chapter 5 discusses the commonsense law of inertia. Chapter 6 discusses the indirect effects of events. Chapter 7 discusses continuous change. Chapter 8 discusses concurrent events. Chapter 9 discusses nondeterministic effects of events. Part III deals with important commonsense domains. Chapter 10 presents axiomatizations of relational and metric space, and discusses reasoning about object identity, space, and time. Chapter 11 presents axiomatizations of the mental states of agents, including beliefs, goals, plans, and emotions. Part IV discusses default reasoning. Part V deals with programs and applications. Chapter 13 discusses the Discrete Event Calculus Reasoner program used to solve event calculus reasoning problems, Chapter 14 discusses several real-world applications, and Chapter 15 discusses the use of answer set programming for commonsense reasoning. Part VI reviews logical and nonlogical methods for commonsense reasoning and discusses their relationship to the event calculus. Chapter 16 reviews logical methods, Chapter 17 reviews nonlogical methods, and Chapter 18 discusses the use of unstructured information for commonsense reasoning. Part VII discusses the acquisition of commonsense knowledge. Part VIII presents my conclusions. Material covered
The skills that make up human commonsense reasoning are complex, and the body of research related to it is large. Because no book can realistically cover every aspect of commonsense reasoning, a choice had to be made about what this book would cover. The coverage of this book was determined by the following considerations. Most instances of commonsense reasoning involve action and change because action and change are pervasive aspects of the world. It is crucial for any method for commonsense reasoning to deal with action and change. Therefore, a large part of this book is devoted to this topic. In addition to reasoning about action and change, or the domain of time, this book covers two other significant domains of commonsense reasoning: space and mental states, including emotions, goals, plans, and beliefs. This book also covers default reasoning and reasoning about object identity. Over the last few decades, researchers have developed a number of logics for commonsense reasoning. It would take much time and space to cover all of these in detail. Hence, this book concentrates on one logic, the event calculus, which incorporates many of the features of the other logics. The reader who understands the event calculus will be well equipped to understand the other logics. They are closely related to the event calculus, and some are provably equivalent to the event calculus—see Appendices D and E. Chapter 16 compares the event calculus with other logics, and detailed Bibliographic notes throughout the book discuss research performed using other logics. Several types of commonsense reasoning are not covered by this book. Common-sense reasoning using probability theory is not covered because this is not a well-developed area. But this book does cover nondeterminism, and some work on the use of probability theory for commonsense reasoning is reviewed (in Section 17.3). Although this book covers most features of the event calculus, it does not cover continuous change described using differential equations; this book does, however, cover continuous change described by closed-form expressions. Supplemental materials
Web site and reasoning programs
The book web site at decreasoner.sourceforge.net contains additional material related to this book. This includes links to tools that can be downloaded, such as the Discrete Event Calculus Reasoner program discussed in Chapter 13 and answer set programming tools discussed in Chapter 15. Exercises and solutions
Exercises are provided at the end of Chapters 2 through 19. Solutions to selected exercises are provided in Appendix F. Solutions to further exercises are available online to instructors who have adopted this text. Register at www.textbooks.elsevier.com for access. Text and figure acknowledgments
Portions of the book, The Snowman, by Raymond Briggs, courtesy of Random House. Portions of the book, Solving the Frame Problem, by Murray Shanahan, courtesy of MIT Press. Portions of the article “A Logic of Actions” by Patrick J. Hayes, in Machine Intelligence, Vol. 6, courtesy of Edinburgh University Press. Portions of the article, “Event Calculus Reasoning Through Satisfiability” by Erik T. Mueller, in Journal of Logic and Computation, courtesy...