Wolfgang Faber

Efficient Computation of the Well-Founded Semantics over Big Data

Data originating from theWeb, sensor readings and social media result in increasingly huge datasets. The so called Big Data comes with new scientific and technological challenges while creating new opportunities, hence the increasing interest in academia and industry. Traditionally, logic programming has focused on complex knowledge structures/programs, so the question arises whether and how it can work in the face of Big Data. In this paper, we examine how the well-founded semantics can process huge amounts of data through mass parallelization. More specifically, we propose and evaluate a parallel approach using the MapReduce framework. Our experimental results indicate that our approach is scalable and that wellfounded semantics can be applied to billions of facts. To the best of our knowledge, this is the first work that addresses large scale nonmonotonic reasoning without the restriction of stratification for predicates of arbitrary arity.

Logic Programming and Nonmotonic Reasoning

I report about a particular approach to heterogenous agent systems, IMPACT, which is strongly related to computational logic. The underlying methods and techniques stem from both non-monotonic reasoning and logic programming. I present three recent extensions to illustrate the generality and usefulness of the approach: (1) incorporating planning, (2) uncertain (probabilistic) reasoning, and (3) reducing the load of serving multiple requests. While (1) illustrates how easy it is to incorporate hierachical task networks into IMPACT, (2) makes heavily use of annotated logic programming and (3) is strongly related to classical first-order reasoning. This paper is a high-level description of (1)–(3), More detailed expositions can be found in [1,2,3,4] from which most parts of this paper are taken. 1 The Basic Framework The IMPACT project (http://www.cs.umd.edu/projects/impact) aims at developing a powerful multi agent system, which (1) is able to deal with heterogenous and distributed data, (2) can be realized on top of arbitrary legacy code, but yet (3) is built on a clear foundational bases and (4) scales up for realistic applications. In this article I am pointing to some recent extensions of the basic framework (which has been implemented and is running) that show very clearly the strong links to computational logic, even though IMPACT’s implementation is not realized on top of a logic related procedural mechanism. To get a bird’s eye view of IMPACT, here are the most important features: – Each IMPACT agent has certain actions available. Agents act in their environment according to their agent program and a well defined semantics determining which of the actions the agent should execute. – Each agent continually undergoes the following cycle: The work I am reporting has been done with many colleagues, notably Th. Eiter, S. Kraus, K. Munoz-Avila, M. Nanni, D. Nau, F. Özcan, T.J. Rogers, R. Ross and, last but not least, V.S Subrahmanian. It resulted in a variety of papers and I gratefully acknowledge their support. T. Eiter, W. Faber, and M. Truszczyński (Eds.): LPNMR 2001, LNAI 2173, pp. 1–21, 2001. c

Supportedly Stable Answer Sets for Logic Programs with Generalized Atoms

Answer Set Programming (ASP) is logic programming under the stable model or answer set semantics. During the last decade, this paradigm has seen several extensions by generalizing the notion of atom used in these programs. Among these, there are dl-atoms, aggregate atoms, HEX atoms, generalized quantifiers, and abstract constraints. In this paper we refer to these constructs collectively as generalized atoms. The idea common to all of these constructs is that their satisfaction depends on the truth values of a set of (non-generalized) atoms, rather than the truth value of a single (non-generalized) atom. Motivated by several examples, we argue that for some of the more intricate generalized atoms, the previously suggested semantics provide unintuitive results and provide an alternative semantics, which we call supportedly stable or SFLP answer sets. We show that it is equivalent to the major previously proposed semantics for programs with convex generalized atoms, and that it in general admits more intended models than other semantics in the presence of non-convex generalized atoms. We show that the complexity of supportedly stable answer sets is on the second level of the polynomial hierarchy, similar to previous proposals and to answer sets of disjunctive logic programs.

The ASP System DLV: Advancements and Applications

We briefly describe the answer set programming system DLV, focusing on some of its peculiar features and mentioning a number of successful applications.

Effectively Solving NP-SPEC Encodings by Translation to ASP

NP-SPEC is a language for specifying problems in NP in a declarative way. Despite the fact that the semantics of the language was given by referring to Datalog with circumscription, which is very close to answer set programming (ASP), so far the only existing implementations are by means of Prolog and via Boolean satisfiability solvers. In this paper, we present translations from NP-SPEC to ASP, and provide an experimental evaluation of existing implementations and the proposed translations into ASP using various ASP solvers. The results show that translating into ASP clearly has an edge over the existing translation into SAT, which involves an intrinsic grounding process. We also argue that it might be useful to incorporate certain language constructs of NP-SPEC into mainstream ASP.

Web reasoning and rule systems

This special issue of the Semantic Web journal is a collection of contributions on the major topics discussed at the Seventh International Conference on Web Reasoning and Rule Systems (RR 2013). These topics include original research from all areas of Web Reasoning, with an emphasis on combinations with Rule Systems. Combining Web Reasoning and Rule Systems embodies the main challenges to be faced by large-scale information systems that deal with heterogeneous, distributed, uncertain, changing, incomplete, and possibly contradictory information. The formalisms and systems need to provide expressiveness, yet be scalable; they need to be aware of context, yet be general; they need to cope with noise, yet provide crisp conclusions. These are just a few examples of the tensions that have to be resolved, and the International Conference on Web Reasoning and Rule Systems (RR) provides a major annual forum for discussions in these areas. The seventh RR conference was held in Mannheim, Germany, during July 26–28, 2013. The conference program included presentations of 23 (full and short) papers that were selected out of 34 submissions. Each submission received at least 3 reviews. The present special issue has been targeted to authors of papers presented at RR 2013 for preparing revised and extended versions of their conference contributions to be considered for publication in the Semantic Web Journal, but it was open to all submissions of papers addressing the conference topics. We received seven submissions, two of which were extended versions of RR 2013 papers. Apart from one submission, which was rejected in pre-screening, each submission underwent a rigorous review process, involving at least three reviews per paper and two review rounds. Four submissions were eventually selected for this special issue, whereas one paper was rejected and another one was withdrawn.