Search

Recent knowledge extraction methods are moving towards ternary and higher-arity relations to capture more information about binary facts. An example is to include the time, the location, and the duration of a specific fact. These relations can be even more complex to extract in advanced domains such as news, where events typically come with different facets including reasons, consequences, purposes, involved parties, and related events. The main challenge consists in first finding the set of facets related to each fact, and second tagging those facets to the relevant category. In this paper, we tackle the above problems by proposing StuffIE, a fine-grained information extraction approach which is facet-centric. We exploit the Stanford dependency parsing enhanced by lexical databases such as WordNet to extract nested triple relations. Then, we exploit the syntactical dependencies to semantically tag facets using distant learning based on Oxford dictionary. We have tested the accuracy of the extracted facets and their semantic tags using DUC'04 dataset. The results show the high accuracy and coverage of our approach with respect to ClausIE, OLLIE, SEMAFOR SRL and Illinois SRL.

The collaborative knowledge base Wikidata is the central storage of Wikimedia projects, containing over 45 million data items. It acts as the hub for interlinking Wikipedia pages about a specific item in different languages, automates features such as infoboxes in Wikipedia, and is increasingly used for other applications such as data enrichment and question answering. Tracking the quality of Wikidata is an important issue for this project. In this paper we focus particularly on the completeness aspect. Several automated techniques have been adopted by Wikis to track and manage completeness, yet these techniques are generally subjective and do not provide a clear quality estimate at the level of entities. In this paper, we present an approach towards measuring Relative Completeness in Wikidata by comparison with data present for similar entities. This relative completeness approach is easily scalable with the introduction of new classes in the knowledge base, and has been implemented for all available entities in Wikidata. The results provide an intuition on the completeness of an entity comparing it with other similar entities. Here, we present our implementation approach along with a discussion on strategies and open challenges.

Industry today employs rule-based diagnostic systems to minimize the maintenance cost and downtime of equipment. Rules are typically used to process signals from sensors installed in equipment by filtering, aggregating, and combining sequences of time-stamped measurements recorded by the sensors. Such rules are often data-dependent in the sense that they rely on specific characteristics of individual sensors and equipment. This dependence poses significant challenges in rule authoring, reuse, and maintenance by engineers especially when the rules require domain knowledge. In this work we propose an approach to address these problems by relying on the well-known Ontology-Based Data Access approach: we propose to use ontologies to mediate the sensor signals and the rules. To this end, we propose a semantic rule language, SDRL, where signals are first class citizens. Our language offers a balance of expressive power, usability, and efficiency: it captures most of Siemens data-driven diagnostic rules, significantly simplifies authoring of diagnostic tasks, and allows to efficiently rewrite semantic rules from ontologies to data and execute over data. We implemented our approach in a semantic diagnostic system and evaluated it. For evaluation we developed a use case of rail systems at Siemens and conducted experiments to demonstrate both usability and efficiency of our solution.

Smart factories are one of the biggest trends in modern manufacturing, also known as Industry 4.0. They reach a new level of process automation and make heavy use of sensors in manufactoring equipment, which brings new challenges to monitoring and diagnostics at smart factories. We propose to address the challenges with a novel rule-based monitoring and diagnostics language that relies on ontologies and reasoning and allows one to write diagnostic tasks at a high level of abstraction. We show that our approach speeds up the diagnostic routine of engineers at Siemens: they can formulate and deploy diagnostic tasks in factories faster than with existing Siemens data-driven solutions. Moreover we show that our diagnostic language, despite the built-in reasoning, allows for efficient execution of diagnostic tasks over large volumes of industrial data. Finally, we implemented our ideas in a prototypical diagnostic system for smart factories.

(a) Situation faced: Frener and Reifer (F&R) is a leader in engineering, fabricating, and installing facades with non-standard designs. The company was looking for comprehensive, domain-specific approaches to improve the company’s control over facade processes, from design to execution and monitoring. What makes process management particularly challenging in this setting are some peculiarities of the domain, such as high levels of variability, unpredictability, and inter-organizational synchronization (vom Brocke et al., BPM Trends, 1, 2015), as well as the non-standard and non-repetitive nature of the designs, which complicates the ability to formulate reliable estimates. Indeed, in many cases the installation department exceeded the number of hours that were initially estimated. (b) Action taken: A group of researchers developed a domain-specific methodology, called PRECISE, that provides methods with which to support the process lifecycle (Dumas et al., Fundamentals of business process management. Springer, 2013) in construction. FR (ii) process implementation, which involves defining short-term (i.e., daily or weekly) schedules for tasks based on actual data on the progress of the work; and (iii) continuous monitoring and measurement of the progress of the work on site. (c) Results achieved: By applying the methodology, which supports a detailed modelling and monitoring of the activities, F&R could perform reliable estimates of progress on tasks and expected cost to completion. For instance, F&R recognized that the budget it had initially estimated was too tight. By analyzing the up-to-date data on the progress of the work and consulting with the workers on the construction site, the company could identify problems and sources of delay promptly and act to mitigate their effects. During the application of PRECISE, F&R recorded an increase in productivity that was estimated to have saved 400 man hours. (d) Lessons learned: Application of the methodology singled out some aspects of the process that should be addressed to improve process management. Flexibility, which is required in dealing with the domain variability, is achieved by defining a process model and a short-term schedule, while the availability of reliable and up-to-date data on the progress of the work is obtained by applying continuous, detailed process monitoring. Engagement of the workers in the process management allows the project to benefit from their expertise (Rosemann and vom Brocke, Handbook on business process management, introduction, methods, and information systems. Springer, 2015), which is the basis of the collaborative approach. However, better IT support for the methodology is needed (Rosemann and vom Brocke, Handbook on business process management, introduction, methods, and information systems. Springer, 2015; Dumas et al., Fundamentals of business process management. Springer, 2013).

In knowledge bases such as Wikidata, it is possible to assert a large set of properties for entities, ranging from generic ones such as name and place of birth to highly profession-specific or background-specific ones such as doctoral advisor or medical condition. Determining a preference or ranking in this large set is a challenge in tasks such as prioritisation of edits or natural-language generation. Most previous approaches to ranking knowledge base properties are purely data-driven, that is, as we show, mistake frequency for interestingness. In this work, we have developed a human-annotated dataset of 350 preference judgments among pairs of knowledge base properties for fixed entities. From this set, we isolate a subset of pairs for which humans show a high level of agreement (87.5% on average). We show, however, that baseline and state-of-the-art techniques achieve only 61.3% precision in predicting human preferences for this subset. We then develop a technique based on a combination of general frequency, applicability to similar entities and semantic similarity that achieves 74% precision. The preference dataset is available at https://www.kaggle.com/srazniewski/wikidatapropertyranking.

We study the problem of evolution for Knowledge Bases (KBs) expressed in Description Logics (DLs) of the DL-Lite family. DL-Lite is at the basis of OWL 2 QL, one of the tractable fragments of OWL 2, the recently proposed revision of the Web Ontology Language. We propose some fundamental principles that KB evolution should respect. We review known model and formula-based approaches for evolution of propositional theories. We exhibit limitations of a number of model-based approaches: besides the fact that they are either not expressible in DL-Lite or hard to compute, they intrinsically ignore the structural properties of KBs, which leads to undesired properties of KBs resulting from such an evolution. We also examine proposals on update and revision of DL KBs that adopt the model-based approaches and discuss their drawbacks. We show that known formula-based approaches are also not appropriate for DL-Lite evolution, either due to high complexity of computation, or because the result of such an action of evolution is not expressible in DL-Lite. Building upon the insights gained, we propose two novel formula-based approaches that respect our principles and for which evolution is expressible in DL-Lite. For our approaches we also developed polynomial time algorithms to compute evolution of DL-Lite KBs.

International audience; Knowledge bases such as Wikidata, DBpedia, YAGO, or the Google Knowledge Vault collect a vast number of facts about the world. But while quite some facts are known about the world, little is known about how much is unknown. For example, while the knowledge base may tell us that Barack Obama is the father of Malia Obama and Sasha Obama, it does not tell us whether these are all of his children. This is not just an epistemic challenge, but also a practical problem for data producers and consumers. We envision that KBs become annotated with information about their recall on specific topics. We show what such annotations could look like, how they could be obtained, and survey related work.

News websites give their users the opportunity to participate in discussions about published articles, by writing comments. Typically, these comments are unstructured making it hard to understand the flow of user discussions. Thus, there is a need for organizing comments to help users to (1) gain more insights about news topics, and (2) have an easy access to comments that trigger their interests. In this work, we address the above problem by organizing comments around the entities and the aspects they discuss. More specifically, we propose an approach for entity and aspect extraction from user comments through the following contributions. First, we extend traditional Named-Entity Recognition approaches, using coreference resolution and external knowledge bases, to detect more occurrences of entities in comments. Second, we exploit part-of-speech tag, dependency tag, and lexical databases to extract explicit and implicit aspects around discussed entities. Third, we evaluate our entity and aspect extraction approach, on manually annotated data, showing that it highly increases precision and recall compared to baseline approaches.

We have developed and implemented the Relational Grid Monitoring Architecture (R-GMA) as part of the DataGrid project, to provide a flexible information and monitoring service for use by other middleware components and applications. R-GMA presents users with a virtual database and mediates queries posed at this database: users pose queries against a global schema and R-GMA takes responsibility for locating relevant sources and returning an answer. R-GMA’s architecture and mechanisms are general and can be used wherever there is a need for publishing and querying information in a distributed environment. We discuss the requirements, design and implementation of R-GMA as deployed on the DataGrid testbed. We also describe some of the ways in which R-GMA is being used.

A basic feature of Terminological Knowledge Representation Systems is to represent knowledge by means of taxonomies, here called terminologies, and to provide a specialized reasoning engine to do inferences on these structures. The taxonomy is built through a representation language called a concept language (or description logic ), which is given a well-defined set-theoretic semantics. The efficiency of reasoning has often been advocated as a primary motivation for the use of such systems. The main contributions of the paper are: (1) a complexity analysis of concept satisfiability and subsumption for a wide class of concept languages; (2) algorithms for these inferences that comply with the worst-case complexity of the reasoning task they perform.

Structured objects are items with defined properties that are to be represented in a computer system. Research in Knowledge Representation (KR) and in Database Design (DB) has produced languages for describing structured objects. Although different in the particular means for defining properties, both areas share the goal of representing a part of the world in a structured way. Moreover, the rise of object-centred formalisms in the last decade has significantly influenced the convergence of languages.

Biological scientists have made large amounts of data available on the Web, which can be accessed by canned or precomputed queries presented via web forms. To satisfy further information needs, users currently have to have a good understanding of SQL and how the data is stored in the database. While accessing information at the ontological layer seems more appropriate, this poses two challenges: (1) to query data in databases and triple stores through an ontology with little performance overhead, and (2) to provide an intuitive web-based access to users that are not IT experts. To address these issues, we draw upon the theory and technology developed for Ontology-Based Data Access for DL-Lite. With an OWL ontology and the DIG-QuOnto reasoner as building blocks, we have developed an application that allows for graphical ontology browsing, query formulation, and answer retrieval via a Web browser. We have evaluated our system for Web-ONtology baseD Extraction of Relational data (Wonder) with an existing large genomics database about horizontal gene transfer and found that it meets both the scalability and the usability requirements.

This chapter provides an introduction to Description Logics as a formal language for representing knowledge and reasoning about it. It first gives a short overview of the ideas underlying Description Logics. Then it introduces syntax and semantics, covering the basic constructors that are used in systems or have been introduced in the literature, and the way these constructors can be used to build knowledge bases. Finally, it defines the typical inference problems, shows how they are interrelated, and describes different approaches for effectively solving these problems. Some of the topics that are only briefly mentioned in this chapter will be treated in more detail in subsequent chapters.

There is a wide variety of data, both static and streaming, being published on and about computing Grids. However one aspect of Grid data that has received relatively little attention thus far is that of incompleteness. With potentially many data sources available to choose from, each with its own limitations in terms of coverage of data sets and of overall reliability, there are many opportunities for inaccuracy due to incompleteness in the data. In this short paper different types of incompleteness are identified in the context of R-GMA as part of an ongoing research project aimed at finding solutions to some of these.

We present a publish/subscribe framework for integrating data streams published by distributed producers. We introduce the idea of republishers which merge a set of data streams, either from producers or other republishers. The resulting hierarchy of producers and republishers can then be used to answer consumer queries over the streams. We discuss how to compute query plans to create such a hierarchy and the maintenance of these plans when the set of streams changes.

In data stream applications, streams typically arise from a geographically distributed collection of producers and may be queried by consumers, which may be distributed as well. In such a setting, a query can be seen as a subscription asking to be informed of all tuples that satisfy a specific condition. We propose to support the publishing and querying of distributed data streams by a publish/subscribe architecture. To enable such a system to scale to a large number of producers and consumers requires the introduction of republishers which collect together data streams and make the merged stream available. If republishers consume from other republishers, a hierarchy of republishers results. We present a formalism that allows distributed data streams, published by independent stream producers, to be integrated as views on a mediated schema. We use the formalism to develop methods to adapt query plans to changes in the set of available data streams and allow consumers to dynamically change which streams they subscribe to.

QUEST is a system for Querying Semantically Tagged documents on the World-Wide Web. The advent of new markup languages, such as XML, facilitates authoring of Web documents that contain not just html tags for instructing a browser how to view a document, but also contain objects that represent the semantic structure of the document. When such documents become widely available, more powerful methods to access and query information on the Web will be possible. The QUEST system was designed and implemented for querying and manipulating documents written in the markup language ohtml. ohtml combines html and objects of the oem data model. QUEST has several new features. First, QUEST can be used to query a combination of hypertext and object structures. Second, The results of queries are ohtml pages and thus of the same type as the data being queried. Third, QUEST implements a new approach for querying semistructured data that produces meaningful answers even when the input data is incomplete, i.e., when some variables of the query cannot be bound to database values. Finally, the experience of developing and using QUEST for querying semantic documents on the Web can be useful for the design and implementation of query languages for xml. This paper provides an overview of the QUEST system and its components.

Traditionally, the core of a Terminological Knowledge Representation System (TKRS) consists of a TBox or terminology, where concepts are introduced, and an ABox or world description, where facts about individuals are stated in terms of concept memberships. This design has a drawback because in most applications the TBox has to meet two functions at a time: On the one hand—similarly to a database schema—frame-like structures with type information are introduced through primitive concepts and primitive roles; on the other hand, views on the objects in the knowledge base are provided through defined concepts. We propose to account for this conceptual separation by partitioning the TBox into two components for primitive and defined concepts, which we call the schema and the view part. We envision the two parts to differ with respect to the language for concepts, the statements allowed, and the semantics. We argue that this separation achieves more conceptual clarity about the role of primitive and defined concepts and the semantics of terminological cycles. Two case studies show the computational benefits to be gained from the refined architecture.

In many knowledge based systems the application domain is modeled in an object-centered formalism. Research in knowledge acquisition has given evidence that this approach allows one to adequately model the conceptual structures of human experts. However, when a novice user wants to describe a particular task to be solved by such a system he has to be well acquainted with the underlying domain model, and therefore is charged with the burden of making himself familiar with it. We aim at giving automated support to a user in this process, which we call task acquisition.

Concept languages have been studied in order to give a formal account of the basic features of frame-based languages. The focus of research in concept languages was initially on the semantical reconstruction of frame-based systems and the computational complexity of reasoning. More recently, attention has been paid to the formalization of other aspects of frame-based languages, such as non-monotonic reasoning and procedural rules, which are necessary in order to bring concept languages closer to implemented systems. In this paper we discuss the above issues in the framework of concept languages enriched with an epistemic operator. In particular, we show that the epistemic operator both introduces novel features in the language, such as sophisticated query formulation and closed world reasoning, and makes it possible to provide a formal account for some aspects of the existing systems, such as rules and definitions, that cannot be characterized in a standard first-order framework.

Configuration is distinguished by two important aspects: the inherent well-structuredness of the problem description, and the synthetic type of problem solving. Up to now there is quite a number of (more or less) ad hoc approaches to configuration — but we are still missing a comprehensive formal treatment. In order to be adequate a formalization of configuration has to take both aspects into account: the well-structuredness and the synthetic problem solving type. In this paper we introduce a formalisation of configuration, called the Constructive Problem Solving (CPS). CPS formalizes the task to construct for a given specification, which is described as a finite set of logical formulas, a semantical model that satisfies the specification. In this approach, a specification consists of two parts. One part describes the domain, the possible components, and their interdependencies. The other part specifies the particular object that is to be configured. We give a sound CPS calculus, which is complete in the sense, that one can compute solutions to configuration problems if they exist. Our CPS approachhas been implemented in a prototype system as a meta interpreter based on a CLP(R) like system. Our ideas will be illustrated with an example from programmable logic control (PLC) units.

Much of the research on concept languages, which also are called terminological languages, has focused on the computational complexity of subsumption. The intractability results can be divided into two groups. First, it has been shown that extending the basic language jy - with constructs containing some form of logical disjunction leads to co-NP-hard subsumption problems. Secondly, adding negation to jy - makes subsumption PSPACE-complete. The main result of this paper is that extending jy - with unrestricted existential quantification makes subsumption NP-complete. This is the first proof of intractability for a concept language containing, whether explicitly or implicitly, no construct expressing disjunction. Unrestricted existential quantification is therefore, alongside disjunction, a source of computational complexity in concept languages.