Nonmonotonic Reasoning

Nonmonotonic reasoning provides formal methods that enable intelligent systems to operate adequately when faced with incomplete or changing information. In particular, it provides rigorous mechanisms for taking back conclusions that, in the presence of new information, turn out to be wrong and for deriving new, alternative conclusions instead. Nonmonotonic reasoning methods provide rigor similar to that of classical reasoning; they form a base for validation and verification and therefore increase confidence in intelligent systems that work with incomplete and changing information. Following a brief introduction to the concepts of predicate logic that are needed in the subsequent chapters, this book presents an in depth treatment of default logic. Other subjects covered include the major approaches of autoepistemic logic and circumscription, belief revision and its relationship to nonmonotonic inference, and briefly, the stable and well-founded semantics of logic programs.

When I first participated in exploring theories of nonmonotonic reasoning in the late 1970s, I had no idea of the wealth of conceptual and mathematical results that would emerge from those halting first steps. This book by Wiktor Marek and Miroslaw Truszczynski is an elegant treatment of a large body of these results. It provides the first comprehensive treatment of two influen tial nonmonotonic logics - autoepistemic and default logic - and describes a number of surprising and deep unifying relationships between them. It also relates them to various modal logics studied in the philosophical logic litera ture, and provides a thorough treatment of their applications as foundations for logic programming semantics and for truth maintenance systems. It is particularly appropriate that Marek and Truszczynski should have authored this book, since so much of the research that went into these results is due to them. Both authors were trained in the Polish school of logic and they bring to their research and writing the logical insights and sophisticated mathematics that one would expect from such a background. I believe that this book is a splendid example of the intellectual maturity of the field of artificial intelligence, and that it will provide a model of scholarship for us all for many years to come. Ray Reiter Department of Computer Science University of Toronto Toronto, Canada M5S 1A4 and The Canadian Institute for Advanced Research Table of Contents 1 1 Introduction .........

Development and environment problems have reached such alarming proportions that the very survival of humanity is now subject to critical and unprecedented threats. In its latest report, the German Advisory Council on Global Change (WBGU) criticizes Germany's global change research community for its lack of international orientation, its bias towards individual disciplines and for its weaknesses in translating scientific results into a form readily accessible to policymakers. The Council identifies alternatives for restructuring the research landscape, focusing primarily on a new 'Syndrome Approach' for global change research. By applying this tool, scientists can systematically describe and analyze the 'diseases' afflicting the Earth System, and thus elaborate response options.

This is the second in a series of workshops that are bringing together researchers from the theoretical end of both the logic programming and artificial intelligence communities to discuss their mutual interests. This workshop emphasizes the relationship between logic programming and non-monotonic reasoning.Luis' Moniz Pereira is Professor in the Department of Computer Science at the Universidade Nova Lisboa, Portugal. Anil Nerode is Professor and Director of the Mathematical Sciences Institute at Cornell University.Topics include: Stable Semantics. Autoepistemic Logic. Abduction. Implementation Issues. Well-founded Semantics. Truth Maintenance. Probabilistic Theories. Applications. Default Logic. Diagnosis. Complexity and Theory. Handling Inconsistency.

Conditionals are omnipresent, in everyday life as well as in scientific environments; they represent generic knowledge acquired inductively or learned from books. They tie a flexible and highly interrelated network of connections along which reasoning is possible and which can be applied to different situations. Therefore, conditionals are important, but also quite problematic objects in knowledge representation. This book presents a new approach to conditionals which captures their dynamic, non-proportional nature particularly well by considering conditionals as agents shifting possible worlds in order to establish relationships and beliefs. This understanding of conditionals yields a rich theory which makes complex interactions between conditionals transparent and operational. Moreover,it provides a unifying and enhanced framework for knowledge representation, nonmonotonic reasoning, belief revision,and even for knowledge discovery.

The capability to reason in a world full of uncertainties, vagueness and ignorance is what distinguishes humans. This ability to argument in a partially known world is the informal definition of common-sense reasoning. The question how common-sense reasoning is performed occupied humanity since we can think of. Last century this issue reached an immense importance. Especially during the last three decades the study of common-sense reasoning became one of the major research topics in Artificial Intelligence (AI). Several formalisms to capture the mechanism of common-sense reasoning have been proposed so far. This book concentrates on presenting the most important formalisms for common-sense reasoning, and, showing that one of the discussed formalisms serves perfectly to capture the mechanism of common-sense reasoning, since this formalism subsumes all other in this book introduced formalisms dealing with common-sense reasoning.

Many approaches in the field of nonmonotonic and ?commonsense? reasoning are actually different representations of the same basic ideas and constructions. This book gives a logical formalization of the original, explanatory approach to nonmonotonic reasoning. It uses the basic formalism of biconsequence relations, as well as derived systems of default, autoepistemic and causal inference, to cover in a single framework such diverse systems as default logic, autoepistemic and modal nonmonotonic logics, input/output and causal logics, argumentation theory, and semantics of general logic programs with negation as failure. This approach provides a clear separation between logical (monotonic) and nonmonotonic aspects of nonmonotonic reasoning. The separation allows, in particular, to single out the logics underlying modern logic programming and restore thereby the connection between logic programming and logic.