Ilias Tachmazidis

Large-scale parallel stratified defeasible reasoning

We are recently experiencing an unprecedented explosion of available data coming from the Web, sensors readings, scientific databases, government authorities and more. Such datasets could benefit from the introduction of rule sets encoding commonly accepted rules or facts, application- or domain-specific rules, com-monsense knowledge etc. This raises the question of whether, how, and to what extent knowledge representation methods are capable of handling huge amounts of data for these applications. In this paper, we consider inconsistency-tolerant reasoning in the form of defeasible logic, and analyze how parallelization, using the MapReduce framework, can be used to reason with defeasible rules over huge datasets. We extend previous work by dealing with predicates of arbitrary arity, under the assumption of stratification. Moving from unary to multi-arity predicates is a decisive step towards practical applications, e.g. reasoning with linked open (RDF) data. Our experimental results demonstrate that defeasible reasoning with millions of data is performant, and has the potential to scale to billions of facts.

Temporal representation and reasoning in OWL 2

The representation of temporal information has been in the center of intensive research activities over the years in the areas of knowledge representation, databases and more recently, the Semantic Web. The proposed approach extends the existing framework of representing temporal information in ontologies by allowing for representation of concepts evolving in time (referred to as “dynamic” information) and of their properties in terms of qualitative descriptions in addition to quantitative ones (i.e., dates, time instants and intervals). For this purpose, we advocate the use of natural language expressions, such as “before” or “after”, for temporal entities whose exact durations or starting and ending points in time are unknown. Reasoning over all types of temporal information (such as the above) is also an important research problem. The current work addresses all these issues as follows: The representation of dynamic concepts is achieved using the “4D-fluents” or, alternatively, the “N-ary relations” mechanism. Both mechanisms are thoroughly explored and are expanded for representing qualitative and quantitative temporal information in OWL. In turn, temporal information is expressed using either intervals or time instants. Qualitative temporal information representation in particular, is realized using sets of SWRL rules and OWL axioms leading to a sound, complete and tractable reasoning procedure based on path consistency applied on the existing relation sets. Building upon existing Semantic Web standards (OWL), tools and member submissions (SWRL), as well as integrating temporal reasoning support into the proposed representation, are important design features of our approach.

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.

Design and evaluation of small-large outer joins in cloud computing environments

Large-scale analytics is a key application area for data processing and parallel computing research. One of the most common (and challenging) operations in this domain is the join. Though inner join approaches have been extensively evaluated in parallel and distributed systems, there is little published work providing analysis of outer joins, especially in the extremely popular cloud computing environments. A common type of outer join is the small–large outer join, where one relation is relatively small and the other is large. Conventional implementations on this condition, such as one based on hash redistribution, often incur significant network communication, while the duplication-based approaches are complex and inefficient. In this work, we present a new method called DDR (duplication and direct redistribution), which aims to enable efficient small–large outer joins in cloud computing environments while being easy to implement using existing predicates in data processing frameworks. We present the detailed implementation of our approach and evaluate its performance through extensive experiments over the widely used MapReduce and Spark platforms. We show that the proposed method is scalable and can achieve significant performance improvements over the conventional approaches. Compared to the state-of-art method, the DDR algorithm is shown to be easier to implement and can achieve very similar or better performance under different outer join workloads, and thus, can be considered as a new option for current data analysis applications. Moreover, our detailed experimental results also have provided insights of current small–large outer join implementations, thereby allowing system developers to make a more informed choice for their data analysis applications.

Computing the stratified semantics of logic programs over big data through mass parallelization

Increasingly huge amounts of data are published on the Web, and generated from sensors and social media. This Big Data challenge poses new scientific and technological challenges and creates new opportunities - thus the increasing attention 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 stratified semantics of logic programming, equivalent to the well-founded semantics for stratified programs, can process huge amounts of data through mass parallelization. In particular, we propose and evaluate a parallel approach using the MapReduce framework. Our experimental results indicate that our approach is scalable and that stratified semantics of logic programming can be applied to billions of facts.

Large-Scale Reasoning with (Semantic) Data

In this paper, we discuss scalable methods for nonmonotonic rule-based reasoning over Semantic Web Data, using MapReduce. This work is motivated by the recent unparalleled explosion of available data coming from the Web, sensor readings, databases, ontologies and more. Such datasets could benefit from the introduction of rule sets encoding commonly accepted rules or facts, application or domain specific rules, commonsense knowledge etc. This raises the question of whether, how, and to what extent knowledge representation methods are capable of handling huge amounts of data for these applications. Our results indicate that our method shows good scalability properties and is able to handle a benchmark dataset of 1 billion triples, bringing it on par with state-of-the-art methods for monotonic reasoning on the semantic web.

Integrated Representation of Temporal Intervals and Durations for the Semantic Web

Representation of temporal information for the Semantic Web often involves qualitative defined information (i.e., information described using natural language terms such as “before” or “lasts longer than”), since precise dates, times and durations are not always available. A basic aspect of temporal information is duration of intervals, thus embedding duration relations into ontologies along with their semantics and reasoning rules is an important practical issue. This work proposes a new representation for intervals and their durations in ontologies by means of OWL properties and reasoning rules in SWRL embedded into the ontology. The proposed representation is based on the decomposition of Interval Duration calculus relations (INDU) offering a compact representation and a tractable reasoning mechanism. Furthermore, by embedding reasoning rules using SWRL into the ontology, reasoning semantics are an integrated part of the representation which can be easily shared and modified without requiring additional specialized reasoning software.

Massively Parallel Reasoning under the Well-Founded Semantics Using X10

Academia and industry are investigating novel approaches for processing vast amounts of data coming from enterprises, the Web, social media and sensor readings in an area that has come to be known as Big Data. Logic programming has traditionally focused on complex knowledge structures/programs. The question arises whether and how it can be applied in the context of Big Data. In this paper, we study how the well-founded semantics can be computed over huge amounts of data using mass parallelization. Specifically, we propose and evaluate a parallel approach based on the X10 programming language. Our experiments demonstrate that our approach has the ability to process up to 1 billion facts within minutes.

Representing Time and Space for the Semantic Web

Representation of temporal and spatial information for the Semantic Web often involves qualitative defined information (i.e., information described using natural language terms such as “before” or ...

Hypercat RDF: Semantic Enrichment for IoT

The rapid growth of sensor networks and smart devices has led to the generation of an increasing amount of information. Such information typically originates from various sources and is published in different formats. One of the key prerequisites for the Internet of Things (IoT) is interoperability. The Hypercat specification defines a lightweight JSON-based hypermedia catalogue, and is tailored towards the existing needs of industry. In this work, we propose a semantic enrichment of Hypercat, defining an RDF-based catalogue. We propose an ontology that captures the core of the Hypercat RDF specification and provides a mapping mechanism between existing JSON and proposed RDF properties. Finally, we propose a new type of search, called Semantic Search, which allows SPARQL-like queries on top of semantically enriched Hypercat catalogues and discuss how this semantic approach offers advantages over what was previously available.