We present the system LAP (Learning Abductive Programs) that is able to learn abductive logic programs from examples and from a background abductive theory. A new type of induction problem has been defined as an exten...
详细信息
ISBN:
(纸本)3540649581
We present the system LAP (Learning Abductive Programs) that is able to learn abductive logic programs from examples and from a background abductive theory. A new type of induction problem has been defined as an extension of the Inductive logicprogramming framework. In the new problem definition, both the background and the target theories are abductive logic programs and abductive derivability is used as the coverage relation. LAP is based on the basic top-down ILP algorithm that has been suitably extended. In particular, coverage of examples is tested by using the abductive proof procedure defined by Kakas and Mancarella [24]. Assumptions can be made in order to cover positive examples and to avoid the coverage of negative ones. and these assumptions can be used as new training data. LAP can be applied for learning in the presence of incomplete knowledge and for learning exceptions to classification rules.
When learning from very large databases, the reduction of complexity is extremely important. Two extremes of making knowledge discovery in databases (KDD) feasible have been put forward. One extreme is to choose a ver...
详细信息
When learning from very large databases, the reduction of complexity is extremely important. Two extremes of making knowledge discovery in databases (KDD) feasible have been put forward. One extreme is to choose a very simple hypothesis language, thereby being capable of very fast learning on real-world databases. The opposite extreme is to select a small data set, thereby being able to learn very expressive (first-order logic) hypotheses. A multistrategy approach allows one to include most of these advantages and exclude most of the disadvantages. Simpler learning algorithms detect hierarchies which are used to structure the hypothesis space for a more complex learning algorithm. The better structured the hypothesis space is, the better learning can prune away uninteresting or losing hypotheses and the faster it becomes. We have combined inductive logicprogramming (ILP) directly with a relational database management system. The ILP algorithm is controlled in a model-driven way by the user and in a data-driven way by structures that are induced by three simple learning algorithms.
We study the relationship between argumentation (abduct ion) and disjunctive logicprogramming. Based on the paradigm of argumentation, an abductive semantic framework for disjunctive logicprogramming is presented, i...
详细信息
ISBN:
(纸本)3540649581
We study the relationship between argumentation (abduct ion) and disjunctive logicprogramming. Based on the paradigm of argumentation, an abductive semantic framework for disjunctive logicprogramming is presented, in which the disjunctions of negative literals are taken as possible assumptions rather than only negative literals as the case of non-disjunctive logicprogramming. In our framework, three semantics PDH, CDH and WFDH are defined by three kinds of acceptable hypotheses to represent credulous reasoning, moderate reasoning and skeptical reasoning in AI, respectively. On the other hand, our semantic framework could be established in a broader class than that of disjunctive programs (called bi-disjunctive logic programs) and, hence, the corresponding abductive framework is abbreviated as BDAS (Bi-Disjunctive Argumentation-theoretic Semantics). Besides its rich expressive power and nondeterminism, BDAS integrates and naturally extends many key semantics, such as the minimal models, EGCWA, the well-founded model, and the stable models. In particular, a novel and interesting argumentation-theoretic characterization of EGCWA is shown. Thus the framework in this paper does not only provides a new way of performing argumentation (abduction) in disjunctive logicprogramming, but also is a simple, intuitive and unifying semantic framework for disjunctive logicprogramming.
Representing and answering the queries about the dynamic behavior of processes in knowledge base systems has become a challenging research area in the field of logicprogramming and knowledgerepresentation systems. I...
详细信息
Representing and answering the queries about the dynamic behavior of processes in knowledge base systems has become a challenging research area in the field of logicprogramming and knowledgerepresentation systems. In this report, we are going to show how transaction logic can be used to efficiently represent dynamic behavior embedded in different domains. The ability of properly representing state changes in transaction logic enables us to express dynamic behavior of processes in different domains. The use of transaction logic to represent dynamic behavior decreases the size of knowledge bases and the query response time in comparison with other existing approaches. The efficiency of our method along with other features of transaction logic and its theoretical basis makes it an appropriate approach to represent dynamic behavior of processes in various domains.
We propose to use modal logic as a logic for coalgebras and discuss it in view of the work done on coalgebras as a semantics of object-oriented programming. Two approaches are taken: First, standard concepts of modal ...
详细信息
We propose to use modal logic as a logic for coalgebras and discuss it in view of the work done on coalgebras as a semantics of object-oriented programming. Two approaches are taken: First, standard concepts of modal logic are applied to coalgebras. For a certain kind of functor it is shown that the logic exactly captures the notion of bisimulation and a complete calculus is given. Examples of verifications of object properties are given. Second, we discuss the relationship of this approach with the coalgebraic logic of Moss (Coalgebraic logic, Ann. pure Appl. logic 96 (1999) 277-317.). (C) 2001 Elsevier Science B.V. All rights reserved.
We extend Answer Set programming with, possibly infinite, open domains. Since this leads, in general, to undecidable reasoning, we restrict the syntax of programs, while carefully guarding useful knowledge representat...
详细信息
ISBN:
(纸本)3540238425
We extend Answer Set programming with, possibly infinite, open domains. Since this leads, in general, to undecidable reasoning, we restrict the syntax of programs, while carefully guarding useful knowledgerepresentation mechanisms such as negation as failure and inequalities. Reasoning with the resulting Conceptual logic Programs can be reduced to finite, normal Answer Set programming, for which reasoners are available. We argue that Conceptual logicprogramming is a useful tool for uniformly representing and reasoning with both ontologies and rules on the Semantic Web, as they can capture a large fragment of the OWL DL ontology language, while extending it in various aspects.
In this overview Ne show how knowledgerepresentation (KR) can be done with the help of generalized logic programs. We start by introducing the core of PROLOG, which is based on definite logic programs. Although this ...
详细信息
ISBN:
(纸本)3540649581
In this overview Ne show how knowledgerepresentation (KR) can be done with the help of generalized logic programs. We start by introducing the core of PROLOG, which is based on definite logic programs. Although this class is very restricted (and will be enriched by various additional features in the rest of the paper), it has a very nice property for KR-tasks: there exist efficient Query-answering procedures - a rop-Down approach and a Bottom-Up evaluation. In addition we can not only handle ground queries but also queries with variables and compute answer-substitutions. It turns out that more advanced KR-tasks can not be properly handled with definite programs. Therefore we extend this basic class of programs by additional features like Negation-as-Finite-Failure, Default-Negation, Explicit Negation, Preferences, and Disjunction. The need for these extensions is motivated by suitable examples and the corresponding semantics are discussed in detail. Clearly, the more expressive the respective class of programs under a certain semantics is, the less efficient are potential Query-answering methods. This point will be illustrated and discussed for every extension. By well-known recursion-theoretic results, it is obvious that there do not exist complete Query-answering procedures for the general case where variables and function symbols are allowed. Nevertheless we consider it an important topic of further research to extract feasible classes of programs where answer-substitutions can be computed.
We investigate the application of abductive logicprogramming, an existing framework for knowledgerepresentation and reasoning, for specifying the knowledge and behaviour of software agents that need to access resour...
详细信息
ISBN:
(纸本)3540140093
We investigate the application of abductive logicprogramming, an existing framework for knowledgerepresentation and reasoning, for specifying the knowledge and behaviour of software agents that need to access resources in a global computing environment. The framework allows agents that need resources to join artificial societies where those resources are available. We show how to endow agents with the capability of becoming and ceasing to be members of societies, for different categories of artificial agent societies, and of requesting and being given or denied resources within societies. The strength of our formulation lies in combining the modelling and the computational properties of abductive logicprogramming for dealing with the issues arising in resource access within artificial agent societies.
Controlled natural languages are subsets of natural languages that can be used to describe a problem in a very precise way, furthermore they can often be translated automatically into a formal notation. We investigate...
详细信息
ISBN:
(纸本)9783642326127
Controlled natural languages are subsets of natural languages that can be used to describe a problem in a very precise way, furthermore they can often be translated automatically into a formal notation. We investigate in this paper how a controlled natural language can be used as a specification language for Answer Set programming (ASP). ASP is a declarative approach to problem solving and has its roots in knowledgerepresentation, logicprogramming, and constraint satisfaction. Solutions of ASP programs are stable models (= answer sets) that build the starting point for question answering. As a proof of concept, we translate a problem specification written in controlled natural language into an ASP program and compute a stable model that contains the answers to a number of questions.
The proceedings contain 23 papers. The special focus in this conference is on Meta-programming in logic. The topics include: A vade-mecum;metaprogramming through intensional deduction;an autoepistemic analysis of meta...
ISBN:
(纸本)9783540562825
The proceedings contain 23 papers. The special focus in this conference is on Meta-programming in logic. The topics include: A vade-mecum;metaprogramming through intensional deduction;an autoepistemic analysis of metalevel reasoning in logicprogramming;an introduction to partial deduction;tutorial on termination of logic programs;definable naming relations in meta-level systems;meta for modularising logicprogramming;an application of meta-level prolog primitives;reflective agents in metalogicprogramming;logic meta-programming facilities in 'LOG the pandora deadlock handler meta-level relation;an experiment;a sensible least herbrand semantics for untyped vanilla meta-programming and its extension to a limited form of amalgamation;a complete resolution method for logical meta-programming languages;model theoretic semantics for demo;intuitions, proof theory and semantics;negation and control in automatically generated logic programs;transforming normal programs by replacement;meta-programming for reordering literals in deductive databases;a new operation in a framework for abstract interpretation of logic programs;representation of fragmentary multilayered knowledge and metaprograms for change, assumptions, objects, and inheritance.
暂无评论