Günther Charwat

Methods for solving reasoning problems in abstract argumentation - A survey

Within the last decade, abstract argumentation has emerged as a central field in Artificial Intelligence. Besides providing a core formalism for many advanced argumentation systems, abstract argumentation has also served to capture several non-monotonic logics and other AI related principles. Although the idea of abstract argumentation is appealingly simple, several reasoning problems in this formalism exhibit high computational complexity. This calls for advanced techniques when it comes to implementation issues, a challenge which has been recently faced from different angles. In this survey, we give an overview on different methods for solving reasoning problems in abstract argumentation and compare their particular features. Moreover, we highlight available state-of-the-art systems for abstract argumentation, which put these methods to practice.

The Fourth Answer Set Programming Competition: Preliminary Report

Answer Set Programming is a well-established paradigm of declarative programming in close relationship with other declarative formalisms such as SAT Modulo Theories, Constraint Handling Rules, PDDL and many others. Since its first informal editions, ASP systems are compared in the nowadays customary ASP Competition. The fourth ASP Competition, held in 2012/2013, is the sequel to previous editions and it was jointly organized by University of Calabria Italy and the Vienna University of Technology Austria. Participants competed on a selected collection of benchmark problems, taken from a variety of research areas and real world applications. The Competition featured two tracks: the Model& Solve Track, held on an open problem encoding, on an open language basis, and open to any kind of system based on a declarative specification paradigm; and the System Track, held on the basis of fixed, public problem encodings, written in a standard ASP language.

The D-FLAT System for Dynamic Programming on Tree Decompositions

Complex reasoning problems over large amounts of data pose a great challenge for computer science. To overcome the obstacle of high computational complexity, exploiting structure by means of tree decompositions has proved to be effective in many cases. However, the implementation of suitable efficient algorithms is often tedious. D-FLAT is a software system that combines the logic programming language Answer Set Programming with problem solving on tree decompositions and can serve as a rapid prototyping tool for such algorithms. Since we initially proposed D-FLAT, we have made major changes to the system, improving its range of applicability and its usability. In this paper, we present the system resulting from these efforts.

D-FLAT^2: Subset Minimization in Dynamic Programming on Tree Decompositions Made Easy

Many problems from the area of AI have been shown tractable for bounded treewidth. In order to put such results into practice, quite involved dynamic programming (DP) algorithms on tree decompositions have to be designed and implemented. These algorithms typically show recurring patterns that call for tasks like subset-minimization. In this paper we present D-FLATˆ2, a system that allows one to obtain DP algorithms (specified in ASP) from simpler principles, where the DP formalization of subset-minimization is performed automatically. We illustrate the method at work by providing several DP algorithms – given in form of ASP programs – that are more space-efficient than existing solutions, while featuring improved readability, reuse and therefore maintainability of ASP code. Experiments show that our approach also yields a significant improvement in runtime performance.

VCWC: A Versioning Competition Workflow Compiler

System competitions evaluate solvers and compare state-of-the-art implementations on benchmark sets in a dedicated and controlled computing environment, usually comprising of multiple machines. Recent initiatives such as [6] aim at establishing best practices in computer science evaluations, especially identifying measures to be taken for ensuring repeatability, excluding common pitfalls, and introducing appropriate tools. For instance, Asparagus [1] focusses on maintaining benchmarks and instances thereof. Other known tools such as Runlim http://fmv.jku.at/runlim/ and Runsolver [12] help to limit resources and measure CPU time and memory usage of solver runs. Other systems are tailored at specific needs of specific communities: the not publicly accessible ASP Competition evaluation platform for the 3rd ASP Competition 2011 [4] implements a framework for running a ASP competition. Another more general platform is StarExec [13], which aims at providing a generic framework for competition maintainers. The last two systems are similar in spirit, but each have restrictions that reduce the possibility of general usage: the StarExec platform does not provide support for generic solver input and has no scripting support, while the ASP Competition evaluation platform has no support for fault-tolerant execution of instance runs.Moreover, benchmark statistics and ranking can only be computed after all solver runs for all benchmark instances have been completed.

Efficient Problem Solving on Tree Decompositions Using Binary Decision Diagrams

Dynamic programming (DP) on tree decompositions is a well studied approach for solving hard problems efficiently. Usually, implementations rely on tables for storing information, and algorithms specify how tuples are manipulated during traversal of the decomposition. However, a bottleneck of such table-based algorithms is relatively high memory consumption. Binary Decision Diagrams (BDDs) and related concepts have been shown to be very well suited to store information efficiently. While several techniques have been proposed that combine DP with efficient BDD-based storage for some particular problems, in this work we present a general approach where DP algorithms are specified on a logical level in form of set-based formula manipulation operations that are executed directly on the BDD data structure. In the paper, we provide several case studies in order to illustrate the method at work, and report on preliminary experiments. These show promising results, both with respect to memory and run-time.

Democratix: A Declarative Approach to Winner Determination

Computing the winners of an election is an important task in voting and preference aggregation. The declarative nature of answer-set programming ASP and the performance of state-of-the-art solvers render ASP very well-suited to tackle this problem. In this work we present a novel, reduction-based approach for a variety of voting rules, ranging from tractable cases to problems harder than NP. The encoded voting rules are put together in the extensible tool Democratix, which handles the computation of winners and is also available as a web application. To learn more about the capabilities and limits of the approach, the encodings are evaluated thoroughly on real-world data as well as oni?źrandomi?źinstances.

ARVis: Visualizing Relations between Answer Sets

Answer set programming ASP is nowadays one of the most popular modeling languages in the areas of Knowledge Representation and Artificial Intelligence. Hereby one represents the problem at hand in such a way that each model of the ASP program corresponds to one solution of the original problem. In recent years, several tools which support the user in developing ASP applications have been introduced. However, explicit treatment of one of the main aspects of ASP, multiple solutions, has received less attention within these tools. In this work, we present a novel system to visualize relations between answer sets of a given program. The core idea of the system is that the user specifies the concept of a relation by an ASP program itself. This yields a highly flexible system that suggests potential applications beyond development environments, e.g., applications in the field of abduction, which we will discuss in a case study.