Standard debugging techniques, like sequential tracing, fail in answer set programming due to its purely declarative approach. We address this problem by means of the graph-oriented computational model underlying the ...
详细信息
ISBN:
(数字)9783540302278
ISBN:
(纸本)3540232427
Standard debugging techniques, like sequential tracing, fail in answer set programming due to its purely declarative approach. We address this problem by means of the graph-oriented computational model underlying the noMoRe system. Although this is no generic solution, it offers a way to make the computation of answer sets transparent within the noMoRe framwork. Apart from the visualization of answer sets in terms of their generating rules, the computation can be animated in different ways.
We consider the simplification of logic programs under the stable-model semantics, with respect to the notions of strong and uniform equivalence between logic programs, respectively. Both notions have recently been co...
详细信息
ISBN:
(纸本)354020721X
We consider the simplification of logic programs under the stable-model semantics, with respect to the notions of strong and uniform equivalence between logic programs, respectively. Both notions have recently been considered for nonmonotonic logic programs (the latter dates back to the 1980s, though) and provide semantic foundations for optimizing programs with input. Extending previous work, we investigate syntactic and semantic rules for program transformation, based on proper notions of consequence. We furthermore provide encodings of these notions in answer-set programming, and give characterizations of programs which are semantically equivalent to positive and Horn programs, respectively. Finally, we investigate the complexity of program simplification and determining semantical equivalence, showing that the problems range between coNP and II2P complexity, and we present some tractable cases.
We aim to specify program transformations in a declarative style, and then to generate executable program transformers from such specifications. Many transformations require non-trivial program analysis to check their...
详细信息
ISBN:
(纸本)9781581137293
We aim to specify program transformations in a declarative style, and then to generate executable program transformers from such specifications. Many transformations require non-trivial program analysis to check their applicability, and it is prohibitively expensive to re-run such analyses after each transformation. It is desirable, therefore, that the analysis information is incrementally updated. We achieve this by drawing on two pieces of previous work: first, Bernhard Steffen's proposal to use model checking for certain analysis problems, and second, John Conway's theory of language factors. The first allows the neat specification of transformations, while the second opens the way for an incremental implementation. The two ideas are linked by using regular patterns instead of Steffen's modal logic: these patterns can be viewed as queries on the set of program paths.
An approach where stable models of Disjunctive logic Programs (DLPs) are computed using SMODELS as a core engine is discussed. The approach is based on two program transformations using which the key tasks in computin...
详细信息
ISBN:
(纸本)354020721X
An approach where stable models of Disjunctive logic Programs (DLPs) are computed using SMODELS as a core engine is discussed. The approach is based on two program transformations using which the key tasks in computing disjunctive stable models can be reduced to computing stable models for normal programs. The implementation is based on an architecture where two SMODELS search engines interact. One of the engines is responsible for generating model candidates for a DLP given as input and the other checks for the minimality of the candidates.
Disjunctive logic programming (DLP) under the answer set semantics is an advanced formalism for knowledge representation and reasoning. It is generally considered more expressive than normal (disjunction-free) logic P...
详细信息
ISBN:
(纸本)354020721X
Disjunctive logic programming (DLP) under the answer set semantics is an advanced formalism for knowledge representation and reasoning. It is generally considered more expressive than normal (disjunction-free) logic programming, whose expressiveness is limited to properties decidable in NP. However, this higher expressiveness comes at a computational cost, and while there are several efficient systems for the normal case, we are only aware of few solid implementations for full DLP. In this paper we investigate novel techniques to couple two main modules (a model generator and a model checker) commonly found in DLP systems more tightly. Instead of using the checker only as a boolean oracle, in our approach, upon a failed check it also returns a so-called unfounded set. Intuitively, this set explains why a model candidate is not an answer set, and the generator employs this knowledge to backtrack until that set is no longer unfounded, which is vastly more efficient than employing full-fledged model checks to control backtracking. Furthermore, we invoke the checker not only for actual model candidates, but also on partial models during model generation to possibly prune the search space. We implemented these approaches in DLV, a state-of-the-art implementation of DLP according to recent comparisons, and carried out experiments;tests on hard benchmark instances show a significant speedup of a factor of two and above.
In this paper we introduce the logic asCSL, an extension of continuous stochastic logic (CSL), which provides powerful means to characterise execution paths of action- and state-labelled Markov chains. In asCSL, path ...
详细信息
ISBN:
(纸本)0769520529
In this paper we introduce the logic asCSL, an extension of continuous stochastic logic (CSL), which provides powerful means to characterise execution paths of action- and state-labelled Markov chains. In asCSL, path properties are characterised by regular expressions over actions and state-formulas. Thus, the executability of a path not only depends on the available actions but also on the validity of certain state formulas in intermediate states. Our main result is that the model checking problem for asCSL can be reduced to CSL model checking on a modified Markov chain, which is obtained through a product automaton construction. We provide a case study of a scalable cellular phone system which shows how the logic asCSL and the model checking procedure can be applied in practice.
Geographical information systems are one of the most important application areas of belief revision. Recently, Wurbel and colleagues [32] have applied the so-called "removed sets revision" (RSR) to the probl...
详细信息
ISBN:
(数字)9783540302278
ISBN:
(纸本)3540232427
Geographical information systems are one of the most important application areas of belief revision. Recently, Wurbel and colleagues [32] have applied the so-called "removed sets revision" (RSR) to the problem of assessment of water heights in a flooded valley. The application was partially satisfactory since only a small part of the valley has been handled. This paper goes one step further, and proposes an extension of (RSR) called "Prioritized Removed Sets Revision" (PRSR). We show that (PRSR) performed using answer set programming makes possible to solve a practical revision problem provided by a real application in the framework of geographical information system (GIS). We first show how PRSR can be encoded into a logic program with answer set semantics, we then present an adaptation of the smodels system devoted to efficiently compute the answer sets in order to perform PRSR. The experimental study shows that the answer set programming approach gives better results than previous implementations of RSR and in particular it allows to handle the whole valley. Lastly, some experimental studies comparing our encoding with implementations based on SAT-solvers are also provided.
We describe a Web-services based conversational agent architecture that combines logical inferences from the WordNet knowledge base and Web content extraction using Google's Web-search API. Our agents interact wit...
详细信息
ISBN:
(纸本)1932415467
We describe a Web-services based conversational agent architecture that combines logical inferences from the WordNet knowledge base and Web content extraction using Google's Web-search API. Our agents interact with users over the Web, as voice-enabled animated characters and can reach wireless devices through Instant Messenger protocol adapters.
Recent research in nonmonotonic logic programming focuses on alternative notions of equivalence. In particular, strong and uniform equivalence are both proposed as useful tools to optimize (parts of) a logic program. ...
详细信息
ISBN:
(数字)9783540302278
ISBN:
(纸本)3540232427
Recent research in nonmonotonic logic programming focuses on alternative notions of equivalence. In particular, strong and uniform equivalence are both proposed as useful tools to optimize (parts of) a logic program. More specifically, given a set P of program rules and a possible optimization Q, strong (resp. uniform) equivalence requires that adding any set S of rules (resp. facts) to P and Q simultaneously results in equivalent programs, i.e., P boolean OR S and Q boolean OR S possess the same stable models. However, in practice it is often necessary to relax this condition in such a way, that dedicated internal atoms in P or Q are no longer allowed to occur in the possible extensions S. In this paper, we consider these relativized notions of both uniform and strong equivalence and provide semantical characterizations by generalizing the notions of UE- and SE-modelhood. These new characterizations capture all notions of equivalence including ordinary equivalence in a uniform way. Finally, we analyze the complexity of the introduced equivalence tests for the important classes of normal and disjunctive logic programs. As a by-product, we reduce the tests for relativized equivalences to ordinary equivalence between two programs. These reductions may serve as a basis for implementation.
There are two types of formalization for induction in logic. In descriptive induction, induced hypotheses describe rules with respect to observations with all predicates minimized. In explanatory induction, on the oth...
详细信息
ISBN:
(纸本)3540229418
There are two types of formalization for induction in logic. In descriptive induction, induced hypotheses describe rules with respect to observations with all predicates minimized. In explanatory induction, on the other hand, hypotheses abductively account for observations without any minimization principle. Both inductive methods have strength and weakness, which are complementary to each other. In this work, we unify these two logical approaches. In the proposed framework, not all predicates are minimized but minimality conditions can be flexibly determined as a circumscription policy. Constructing appropriate policies, we can intentionally minimize models of an augmented axiom set. As a result, induced hypotheses can have both conservativeness and explainability, which have been considered incompatible with each other in the literature. We also give two procedures to compute inductive hypotheses in the proposed framework.
暂无评论