Kakas, A.C., van Nu#elen, B., Denecker, M.: A-System: Problem solving through abduction. In Nebel, B., ed.: Proc. 17th IJCAI, Seattle, WA, MKP (2001) 591--596  Home/Search   Document Details and Download   Summary   Related Articles   Check

No context found.

A. C. Kakas, B. van Nu#elen, and M. Denecker. A-System: Problem solving through abduction. In B. Nebel, editor, Proceedings of the 17th International Joint Conference on Artificial Intelligence, pages 591--596, Seattle, Washington, USA, August 2001. Morgan Kaufmann Publishers.

....abduction in the context of ID logic. ID logic was defined in an attempt to cope with the epistemological challenges of logic programming and gives answers to the epistemological questions raised in section 7.2. At the computational level, we are currently developing a system called the A system [63, 118] integrating features of ACLP and SLDNFAC with special attention to the search in the abductive computation. During the computation, the selection and evaluation of choice points is guided by information obtained from a constraint store associated to the abductive solution. With this information ....

A.C Kakas, Bert Van Nu#elen, and Marc Denecker. A-system : Problem solving through abduction. In Proceedings of IJCAI'01 - Seventeenth International Joint Conference on Artificial Intelligence, pp. 591-597, 2001.

....application domains, such as planning, diagnosis, multiagent coordination, However few of the implemented abductive systems were in general successful. Among the most evolved systems were the recent ACLP [2, 3] and SLDNFAC [1, 6] which combine constraint logic programming with abduction. In [4], a new system the A System is presented which adds to these systems a better control strategy, and can be seen as the descendant of both. Especially it adds a propagation phase after each choice the system makes. The result is a more robust and modular system which is capable to solve a wider ....

....control strategy, and can be seen as the descendant of both. Especially it adds a propagation phase after each choice the system makes. The result is a more robust and modular system which is capable to solve a wider range of problems than the older systems. The A System has been evaluated in [4] within two problem domains: simple CSP problems like Nqueens and graph coloring and planning problems like the blocks world and logistics domain. SLDNFAC and ACLP were able to solve the rst type quite easily but they ran into deep control problems on planning instances. Exact on these, the ....

A.C Kakas, Bert Van Nuelen, and Marc Denecker. A-system : Problem solving through abduction. In Proceedings of IJCAI'01 - Seventeenth International Joint Conference on Arti cial Intelligence. International Joint Conferences on Arti cial Intelligence, Inc. and American Association for Arti cial Intelligence, 2001. accepted.

....problems in the context of reasoning on continuous change. The principle of integrating abductive reasoning and constraint solving (in the natural numbers) was described in [25, 24] An extension of SLDNFA that integrates i abduction and nite domain constraint solving was presented in [26]. SLDNFA can be extended along similar lines with constraint solving for reasoning with linear equations and inequalities in the real numbers. A prototype implementation of an integration of SLDNFA and CLP(R) a constraint solver for real number linear arithmetic, was presented in [3] Below, two ....

A.C Kakas, Bert Van Nuelen, and Marc Denecker. A-system : Problem solving through abduction. In Proceedings of IJCAI'01 - Seventeenth International Joint Conference on Arti cial Intelligence. International Joint Conferences on Arti cial Intelligence, Inc. and American Association for Arti cial Intelligence, 2001. 34

....between the several approaches. In preparation of this paper we have formulated the most interesting ones found at [9] in our framework. We will discuss two of them (tweety ies and buying a car) to clarify the framework. All the experiments are done by using the abductive solver Asystem [1, 8] because there exists a relationship between ID Logic and Abductive Logic Programming. ID Logic theories can be transformed to the Abductive Logic Programming framework [ by integrating all the inductive de nitions into one de nition. In the obtained theory, the resulting de nition ....

....corresponds to the Abductive Logic Program, the open predicates in it to the abducibles and the FOL part to the integrity constraints. Because the Asystem semantics is the Well Founded semantics, the answers generated by the Asystem correspond to ID Logic models. Due lack of space we refer to [1, 8] for a detailed explanation of the abductive solver. And all encodings are public available on the homepage of the Asystem [1] 4.1 Example 1: Tweety ies This is the classic example in Logic Programming. In our limited view on biological classi cation we have a general class of birds and two ....

A.C Kakas, Bert Van Nuelen, and Marc Denecker. A-system : Problem solving through abduction. In B. Nebel, editor, Proceedings of the Seventeenth International Joint Conference on Arti cial Intelligence, volume 1, pages 591{ 596. IJCAI, inc and AAAI, Morgan Kaufmann Publishers, Inc, 2001.

....However, an ID Logic theory is also a generalization of standard logic programming. There, the basic computational task is deduction: Given a query Q (with Q a conjunction of literals) the task is to find a substitution a such that Qa is true in the unique model of the program. The Asystem [13, 1] is able to combine both computational tasks, because it is an implementation of the SLDNFA procedure [9] That SLDNFA procedure 5 This is exactly the least Herbrand model in the case of a definite program. is a sound and complete abductire extension of the SLDNF procedure that is at the basis of ....

.... Given an ID Logic theory (Z) r) and a query 4 Q, SLDNFA computes a A and a substitution (if they exist, otherwise it fails) such that (1) the well founded model of is two valued and (2) Q and r are true in the well founded model of (Z)UA) The 4system has been used in a number of experiments [13], including com putationally hard problems like AI planning. 3.3 Extensions of ID Logic In the above discussion, we have formulated a basic version of ID Logic and omitted a number of aspects which are of minor importance for the understanding of this paper. Besides the formulated core, ....

A.C. Kakas, Bert Van Nuffelen, and Marc Denecker. A-system: Problem solving through abduction. In B. Nebel, editor, Proceedings of the Seventeenth International Joint Conference on Artificial Intelligence, pages 591-596. International Joint Conferences on Artificial Intelligence, Inc. and American Association for Artificial Intelligence, 2001.

....analysis that characterizes database repairs in terms of a certain set of models of the inconsistent database (those that, intuitively, minimize the amount of inconsistent information) The other approach is based on abductive reasoning. For this we use an abductive solver (A system, [27]) that implements SLDNFA resolution [16, 17] for computing a list of data facts that should be inserted to the database or retracted from it in order to keep the data consistent. A corresponding application was introduced and described in greater details in [7] Here we review it in order to keep ....

....the open predicates of a theory in ID logic correspond to the abducibles in an abductive logic program. Consequently, solutions of abductive logic programs that are computed by an abductive solver are also models of the corresponding ID logic theory. Here we use such a solver, called the A system [7, 27] for computing solutions. The main idea of this solver is to reduce a high level speci cation into a lower level constraint store, which is managed by a constraint solver. The solver combines the refutation procedures SLDNFA [17] and ACLP [29] and uses an improved control strategy. In our case, ....

T.Kakas, B.Van Nuelen, M.Denecker. A-System: Problem solving through abduction. Proc. IJCAI'01 , 2001.

....Abduction Bert Van Nu elen and Antonis Kakas Dept. Computer Science, K.U.Leuven Celestijnenlaan 200A, Leuven, Belgium bertv cs.kuleuven.ac.be Dept. of Computer Science, University of Cyprus 75 Kallipoleos St. Nicosia, Cyprus antonis ucy.ac. cy 1 General Information The A system [4] is a new system for performing abductive reasoning within the framework of Abductive Logic Programming (ALP) The principles behind the system are founded by work on two earlier systems ACLP [2, 3] and SLDNFA(C) 1, 6] The basic inference mechanism of the system combines abductive logic ....

....process might get stuck in a local satis ability check of this. 4.1 Benchmarks and Comparison Currently very few benchmarks are available for testing the full capabilities of the A system with the AIPS planning competition test set being the best example. The A system has been evaluated in [4] with some of the problems from this set. A comparison of the A system with other systems can be separated in two classes: to other general purpose systems like Smodels and DLV or more specialized systems like AI planners. At this moment no extensive study has been made to compare the A system ....

A.C Kakas, Bert Van Nuelen, and Marc Denecker. A-system : Problem solving through abduction. In Proceedings of IJCAI'01, 2001. to appear.

....on a metatheory relating a number of di erent (possibly inconsistent) input databases with a consistent output database. The underlying language is that of ID logic [9] which can be embedded in an abductive logic program. Our composing system is implemented by the abductive solver the, A system [18]. In the context of this work, we extended this system with an optimizing component that will allow us to compute preferred coherent solutions to restore the consistency of the database. Our approach is related to other work on the use of abduction in the context of databases. 16] proposed to ....

....follows we shall x a preference criterion for choosing the best repairs and omit its notation whenever possible. problem domain (ID logic, 9] an abductive solver that is based on an abductive refutation procedure (SLDNFA, 10] and a computational model for controlling the search (A system [18]) 3.1 ID logic and abductive logic programming ID logic [9] is a framework for declarative knowledge representation that extends classical logic with inductive de nitions. This logic incorporates two types of knowledge: de nitional and assertional. Assertional knowledge is a set of rst order ....

[Article contains additional citation context not shown here]

T.Kakas, B.Van Nuelen, M.Denecker. A-System: Problem solving through abduction. Proc. IJCAI'01, 2001.

No context found.

Kakas, A.C., van Nu#elen, B., Denecker, M.: A-System: Problem solving through abduction. In Nebel, B., ed.: Proc. 17th IJCAI, Seattle, WA, MKP (2001) 591--596

No context found.

Kakas, A.C., van Nu#elen, B., Denecker, M.: A-System: Problem solving through abduction. In Nebel, B., ed.: Proceedings of the 17th IJCAI, Seattle, Washington, USA, Morgan Kaufmann Publishers (2001) 591--596

No context found.

Kakas, A.C., van Nuffelen, B., Denecker, M.: A-System: Problem solving through abduction. In Nebel, B., ed.: Proceedings of the 17th International Joint Conference on Artificial Intelligence, Seattle, Washington, USA, Morgan Kaufmann Publishers (2001) 591--596

No context found.

Kakas, A.C., Nu#elen, B.V., Denecker, M.: A-System: Problem solving through abduction. In Nebel, B., ed.: Proceedings of IJCAI 2001, Seattle, Washington, USA, Morgan Kaufmann (2001) 591--596

No context found.

A. C. Kakas, B. Van Nu#elen, and M. Denecker. A-system: Problem solving through abduction. In Proc. IJCAI-2001. Morgan Kaufmann, 2001.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC