The Semantic Web

Contrastive reasoning is the reasoning with contrasts which are expressed as contrary conjunctions like the word ”but” in natural language. Contrastive answers are more informative for reasoning with inconsistent ontologies, as compared with the usual simple Boolean answer, i.e., either ”yes” or ”no”. In this paper, we propose a method of computing contrastive answers from inconsistent ontologies. The proposed approach has been implemented in the system CRION (Contrastive Reasoning with Inconsistent ONtologies) as a reasoning plug-in in the LarKC (Large Knowledge Collider) platform. We report several experiments in which we apply the CRION system to some realistic ontologies. This evaluation shows that contrastive reasoning is a useful extension to the existing approaches of reasoning with inconsistent ontologies.

In order to support the vision of the Semantic Web, ontology reasoning needs to be highly scalable and efficient. A natural way to achieve scalability and efficiency is to develop parallel ABox reasoning algorithms for tractable OWL 2 profiles to distribute the load between different computation units within a reasoning system. So far there have been some work on parallel ABox reasoning algorithms for the pD* fragment of OWL 2 RL. However, there is still no work on parallel ABox reasoning algorithm for OWL 2 EL, which is the language for many influential ontologies (such as the SNOMED CT ontology). In this paper, we extend a parallel TBox reasoning algorithm [5] for ${\mathcal{ELH_{R+}}}$ to parallel ABox reasoning algorithms for $\mathcal{ELH}_{\bot, \mathcal{R}+}$, which also supports the bottom concept so as to model disjointness and inconsistency. In design of algorithms, we exploit the characteristic of ABox reasonings to improve parallelisation and reduce unnecessary resource cost. Our evaluation shows that a naive implementation of our approach can compute all ABox entailments of a Not-Galen− ontology with about 1 million individuals and 9 million axioms in about 3 minutes.

With the rapid increase of RDF data, the SPARQL query processing has received much attention. Currently, most RDF databases store RDF data in a relational table called triple table and carry out several join operations on the triple tables for SPARQL query processing. However, the execution plans with many joins might be inefficient due to a large amount of intermediate data being passed between join operations. In this paper, we propose a triple filtering method called RP-Filter to reduce the amount of intermediate data. RP-Filter exploits the path information in the query graphs and filters the triples which would not be included in final results in advance of joins. We also suggest an efficient relational operator RFLT which filters triples by means of RP-Filter. Experimental results on synthetic and real-life RDF data show that RP-Filter can reduce the intermediate results effectively and accelerate the SPARQL query processing.

Ontology matching is a crucial task for data integration and management on the Semantic Web. The ontology matching techniques today can solve many problems from heterogeneity of ontologies to some extent. However, for matching large ontologies, most ontology matchers take too long run time and have strong requirements on running environment. Based on the MapReduce framework and the virtual document technique, in this paper, we propose a 3-stage MapReduce-based approach called V-Doc+ for matching large ontologies, which significantly reduces the run time while keeping good precision and recall. Firstly, we establish four MapReduce processes to construct virtual document for each entity (class, property or instance), which consist of a simple process for the descriptions of entities, an iterative process for the descriptions of blank nodes and two processes for exchanging the descriptions with neighbors. Then, we use a word-weight-based partition method to calculate similarities between entities in the corresponding reducers. We report our results from two experiments on an OAEI dataset and a dataset from the biology domain. Its performance is assessed by comparing with existing ontology matchers. Additionally, we show how run time is reduced with increasing the size of cluster.

In an open context such as the Semantic Web, information providers usually rely on different ontologies to semantically characterize contents. In order to enable interoperability at a semantic level, ontologies underlying information sources must be linked by discovering alignments, that is, set of correspondences or mappings. The aim of this paper is to provide a formal model (i.e.,

Semantic Flow Networks

) to represent networks of ontologies and alignments with the aim to investigate the problem of composite mapping discovery.

Semantic Flow Networks

(

SFN

) differ from other models of networks of ontologies for two main aspects.

SFN

consider constraints over mappings that are necessary to take into account their dependencies. Moreover, a different notion of mapping, that is, compound mapping is considered. Complexity results and a CSP formulation for composite mapping discovery are provided.

DBpedia has been proved to be a successful structured knowledge base, and large scale Semantic Web data has been built by using DBpedia as the central interlinking-hubs of the Web of Data in English. But in Chinese, due to the heavily imbalance in size (no more than one tenth) between English and Chinese in Wikipedia, there are few Chinese linked data are published and linked to DBpedia, which hinders the structured knowledge sharing both within Chinese resources and cross-lingual resources. This paper aims at building large scale Chinese structured knowledge base from Hudong, which is one of the largest Chinese Wiki Encyclopedia websites. In this paper, an upper-level ontology schema in Chinese is first learned based on the category system and Infobox information in Hudong. Totally, there are 19542 concepts are inferred, which are organized in hierarchy with maximally 20 levels. 2381 properties with domain and range information are learned according to the attributes in the Hudong Infoboxes. Then, 802593 instances are extracted and described using the concepts and properties in the learned ontology. These extracted instances cover a wide range of things, including persons, organizations, places and so on. Among all the instances, 62679 of them are linked to identical instances in DBpedia. Moreover, the paper provides RDF dump or SPARQL to access the established Chinese knowledge base. The general upper-level ontology and wide coverage makes the knowledge base a valuable Chinese semantic resource. It not only can be used in Chinese linked data building, the fundamental work for building multi lingual knowledge base across heterogeneous resources of different languages, but also can largely facilitate many useful applications of large-scale knowledge base such as knowledge question-answering and semantic search.

In ontological theories, is-a hierarchy must represent the essential property of things and hence should be single-inheritance, since the essential property of things cannot exist in multiple. However, we cannot avoid multiperspective issues when we build an ontology because the user often want to understand things from their own viewpoints. Especially, in the Semantic Web, the variety of users causes the variety of viewpoints to capture target domains. In order to tackle this multi-perspective issue, we adopt an approach of dynamically generating is-a hierarchies according to the viewpoints of users from an ontology using single-inheritance. This article discusses a framework for dynamic is-a hierarchy generation with ontological consideration on is-a hierarchies generated by it. Then, the author shows its implementation as a new function of Hozo and its applications to a medical ontology for dynamically generation of is-a hierarchies of disease. Through the function, users can understand an ontology from a variety of viewpoints. As a result, it could contribute to comprehensive understanding of the ontology and its target world.

The Linking Open Data(LOD) cloud is a collection of linked Resource Description Framework (RDF) data with over 26 billion RDF triples. Consuming linked data is a challenging task because each data set in the LOD cloud has specific ontology schema, and familiarity with ontology schema is required in order to query various linked data sets. However, manually checking each data set is time-consuming, especially when many data sets from various domains are used. This difficulty can be overcome without user interaction by using an automatic method that integrates different ontology schema. In this paper, we propose a Mid-Ontology learning approach that can automatically construct a simple ontology, linking related ontology predicates (class or property) in different data sets. Our Mid-Ontology learning approach consists of three main phases: data collection, predicate grouping, and Mid-Ontology construction. Experimental results show that our Mid-Ontology learning approach successfully integrates diverse ontology schema, and effectively retrieves related information.

In this paper, we propose an ontological formulation of the planning domain and its OWL 2 formalization. The proposed metamodel conceptualizes planning rules and actions and the

causality

between them. We also show that our planning metamodel can be seen as a relevant scenario of the Open Provenance Model (OPM) and we define our planning OPM profile.

This ontological representation is then exploited to define automated means for the verification of correctness and consistency of a planning domain model. We claim that Semantic Web technologies can provide an effective solution to this important – and often underestimated – problem for planning applications.

To perform matchmaking in Web-based scenarios where data are often incomplete, we propose an extended conjunctive query answering (CQA) problem, called

abductive CQA problem

, in Description Logic ontologies. Given a consistent ontology and a conjunctive query, the abductive CQA problem computes all

abductive answers

to the query in the ontology. An abductive answer is an answer to the query in some consistent ontology enlarged from the given one by adding a bounded number of individual assertions, where the individual assertions that can be added are confined by user-specified concept or role names. We also propose a new approach to matchmaking based on the abductive CQA semantics, in which offer information is expressed as individual assertions, request information is expressed as conjunctive queries, and matches for a request are defined as abductive answers to a conjunctive query that expresses the request. We propose a sound and complete method for computing all abductive answers to a conjunctive query in an ontology expressed in the Description Logic Program fragment of OWL 2 DL with the Unique Name Assumption. The feasibility of this method is demonstrated by a real-life application, rental matchmaking, which handles requests for renting houses.

In this paper, we present GeniUS, a generic topic and user modeling library for the Social Semantic Web that enriches the semantics of social data and status messages particularly. Given a stream of messages, it allows for generating topic and user profiles that summarize the stream according to domain- and application-specific needs which can be specified by the requesting party. Therefore, GeniUS can be applied in various application settings. In this paper, we analyze and evaluate GeniUS in six different application domains. Given users’ status messages from Twitter, we investigate the quality of profiles that are generated by different GeniUS user modeling strategies for supporting various recommendation tasks ranging from product recommendations to more specific recommendations as required in book or software product stores. Our evaluation shows that GeniUS succeeds in inferring the semantic meaning of Twitter status messages. We prove that it can successfully adapt to a given domain and application context allowing for tremendous improvements of the recommendation quality when domain-specific semantic filtering is applied to remove noise from the profiles.

Traditionally, Linked Data query engines execute SPARQL queries over a materialised repository which on the one hand, guarantees fast query answering but on the other hand requires time and resource consuming preprocessing steps. In addition, the materialised repositories have to deal with the ongoing challenge of maintaining the index which is – given the size of the Web – practically unfeasible. Thus, the results for a given SPARQL query are potentially out-dated. Recent approaches address the result freshness problem by answering a given query directly over dereferenced query relevant Web documents. Our work investigate the problem of an efficient selection of query relevant sources under this context. As a part of query optimization, source selection tries to estimate the minimum number of sources accessed in order to answer a query. We propose to summarize and index sources based on frequently appearing query graph patterns mined from query logs. We verify the applicability of our approach and empirically show that our approach significantly reduces the number of relevant sources estimated while keeping the overhead low.

Ontology debugging helps users to understand the unsatisfiability of a concept in an ontology by finding minimal unsatisfiability-preserving sub-ontologies (MUPS) of the ontology for the concept. Although existing approaches have shown good performance for some real life ontologies, they are still inefficient to handle ontologies that have many MUPS for an unsatisfiable concept. In this paper, we propose an efficient approach to debugging ontologies based on a set of patterns. Patterns provide general information to explain unsatisfiability but are not dependent on a specific ontology. In this approach, we make use of a set of heuristic strategies and construct a directed graph w.r.t. the hierarchies where the depth-first search strategy can be used to search paths. The experiments show that our approach has gained a significant improvement over the state of the art and can find considerable number of MUPS.

Distributed SPARQL queries enable users to retrieve information by exploiting the increasing amount of linked data being published. However, industrial-strength distributed SPARQL query processing is still at its early stage for efficiently answering queries. Previous research shows that it is possible to apply methods from graph theory to optimize the performance of distributed SPARQL. In this paper we describe a framework that can simulate arbitrary RDF data networks to evaluate different approaches of distributed SPARQL query processing. Using this framework we further explore the graph traversal algorithms for distributed SPARQL optimization. We present an implementation of a Minimum-Spanning-Tree-based (MST-based) algorithm for distributed SPARQL processing, the performance of which is compared to other approaches using this evaluation framework. The contribution of this paper is to show that a MST-based approach seems to perform much better than other non graph-traversal-based approaches, and to provide an evaluation framework for evaluating distributed SPARQL processing.

When searching for RDF vocabularies, users often feel hindered by the lengthy description of a retrieved vocabulary from judging its relevance. A natural strategy for dealing with this issue is to generate a summary of the vocabulary description that compactly carries its main theme and reveals its relevance to the user’s information need. In this paper, we present a new solution to this problem of vocabulary summarization, which has been defined as ranking and selecting RDF sentences in our previous work. Firstly, we propose a novel bipartite graph representation of vocabulary description, on which we carry out a stochastic analysis of a random surfer’s behavior, from which we derive a new centrality measure for RDF sentences called BipRank. Further, we improve it by investigating the patterns of RDF sentences and employing their statistical features. Then, we combine BipRank with query relevance and cohesion metrics into an aggregate objective function to be optimized for the selection of RDF sentences. Our experiments on real-world vocabularies demonstrate the superiority of our approach to the baseline, and also validate its scalability in practice.

Concurrent fine grained updates are essential for using RDF stores in dynamic modern Web applications, where users increasingly contribute content as often as they read content. SPARQL Update is a language proposed by the W3C for fine grained updates for RDF stores. In this work we propose an operational semantics for an update language for RDF, which models core features of SPARQL Update. Firstly, an abstract syntax for RDF and updates is presented. Secondly, the operational semantics is defined using relations over the abstract syntax. The operational semantics specifies all possible operational behaviours of updates in the presence of an RDF store. The specification is useful as a common reference for compiler engineers and as a foundation for the static analysis of updates.

Recognizing diseases from theoretical perspective can help ordinary people have a general understanding of medicine. The usual process of identifying syndromes or diseases in Traditional Chinese Medicine (TCM) is by confirming the frequently symptom patterns. Semantic Web and ontologies introduce well-structured controlled vocabularies for biomedical science. The direct correspondence between symptoms and syndromes can be formatted to semantic inference rules as a additional knowledge upon a medical ontology.

In this paper, we present a simplified rule-based diagnostic system for febrile disease theory in TCM, which make use of the capability of semantic inference based on medical ontology. Actually the method is rather general for logic-based medical diagnosis, and we show that without interpreting clinical data, the medical knowledge itself can be applied to do basic clinical diagnosis.

Measuring semantic relatedness plays an important role in information retrieval and Natural Language Processing. However, little attention has been paid to measuring semantic relatedness between named entities, which is also very significant. As the existing knowledge based approaches have the entity coverage issue and the statistical based approaches have unreliable result to low frequent entities, we propose a more comprehensive approach by leveraging Linked Open Data (LOD) to solve these problems. LOD consists of lots of data sources from different domains and provides rich a priori knowledge about the entities in the world. By exploiting the semantic associations in LOD, we propose a novel algorithm, called LODDO, to measure the semantic relatedness between named entities. The experimental results show the high performance and robustness of our approach.

With more data repositories constantly being published on the Web, choosing appropriate data sources to interlink with newly published datasets becomes a non-trivial problem. It is necessary to choose both the repositories to link to and the relevant subsets of these repositories, which contain potentially matching individuals. In order to do this, detailed information about the content and structure of semantic repositories is often required. However, retrieving and processing such information for a potentially large number of datasets is practically unfeasible. In this paper, we propose an approach which utilises an existing semantic web index in order to identify potentially relevant datasets for interlinking and rank them. Furthermore, we adapt instance-based ontology schema matching to extract relevant subsets of selected data source and, in this way, pre-configure data linking tools.

The continuous growth of the Linked Data Web brings us closer to the original vision of the Web as an interconnected network of machine-readable resources. There is, however, an essential aspect in principle still missing from this vision, i.e., the ability for the Web user to interact directly with the Linked Data in a read/write manner. In this paper we introduce a lifecycle and associated mechanism to enable a domain-agnostic read/write interaction with Linked Data in the context of a single data provider. Our solution uses an ontology to build a binding front-end for a given RDF model, in addition to RDFa to maintain the semantics of the resulting form/widget components. On the processing side, a RESTful Web service is provided to seamlessly manage semantic widgets and their associated data, and hence enable the read/write data interaction mechanism. The evaluation shows that the generation process presents no performance issues, while the content overhead required for the actual form-data binding is kept to a minimum.

RDFa is a syntactic format that allows RDF triples to be integrated into hypertext content of HTML/XHTML documents. Although a growing number of methods or tools have been designed attempting at generating or digesting RDFa, comparatively little work has been carried out on finding a generic solution for publishing existing RDF data sets with the RDFa serialisation format. This paper proposes a generic and lightweight approach to generating semantically-enriched hypertext content by embedding RDF triples derived from diverse provenances in terms of a concept of topic nodes which will be automatically recommended by our discovery algorithm. RDFa

2

is a proof-of-concept implementation for our approach and works as an online platform assisting Web content publishers in semi-automatically generating, personalising and curating pages with RDFa. RDFa

2

has been introduced and employed by students in a master level course and the experimental results as well as additional case studies indicate the validity of this approach to generating triple-embedded Web documents such as online profiles and vocabularies with little user intervention.

Although a formal query language, SPARQL, is available for accessing DBpedia, it remains challenging for users to query the knowledge unless they are familiar with the syntax of SPARQL and the underlying ontology. We have developed both an intuitive

semantic graph

notation or interface allowing one to pose a query by annotating a graph with natural language terms denoting entities and relations and a system that automatically translates the query into SPARQL to produce an answer. Our key contributions are the robust techniques, combining statistical association and semantic similarity, that map user terms to the most appropriate classes and properties used in the DBpedia Ontology.

The National Center for Biomedical Ontology (NCBO) BioPortal provides common access for browsing and querying a large set of ontologies that are commonly used in biomedical communities. One of our missions is to align lexical features (i.e., textual definitions) that are commonly used in these ontologies across different representation formats with standard tags and to represent them in a standard way to the users. The Simple Knowledge Organization System (SKOS) is a recommendation of the World-Wide-Web Consortium (W3C) for a common data model for sharing and linking knowledge organization systems on the Semantic Web. The BioPortal is in the process of adopting SKOS in the backend representation for its content. During this process, we discovered that there exists a set of commonly-used lexical features shared by the biomedical ontologies that SKOS does not yet represent. In this paper, we discuss our proposed SKOS extensions to cover this set of commonly used lexical features, the rationales, and the detailed description of each proposed construct.

In this paper, we introduce a new approach for constructing complex mappings between ontologies by transforming it to a rule learning process. Derived from the classical Inductive Logic Programming, our approach uses instance mappings as training data and employs tailoring heuristics to improve the learning efficiency. Empirical evaluation shows that our generated Horn-rule mappings are meaningful.

For new data supplier who wants to join the web of data club, it’s difficult to find new links between local repository and data sets in the web of data to make local data well-connected or harmonize with other data. The purpose of this research is not for finding similar entities but discovering new potential link for helping users have more choice for using multiple links instead of only using “owl:sameAs”. The approach use information retrieval technique index the data sets and Page Rank and graph theory analyze RDF document to filter links. We implemented our method using Dbpedia data sets and two open ontologies, the results showed our approach can discover new links with highly accuracy.

The concept of contexts is widely used in artificial intelligence. Several recent attempts have been made to formalize multi-context systems (MCS) for ontology applications. However, these approaches are unable to handle probabilistic knowledge. This paper introduces a formal framework for representing and reasoning about uncertainty in multi-context systems (called p-MCS). Some important properties of p-MCS are presented and an algorithm for computing the semantics is developed. Examples are also used to demonstrate the suitability of p-MCS.

The ultimate vision of the semantic web is to enable computers to understand and process the information published on the web. This vision is being primarily achieved by web ontologies which semantically annotate the data. In order to effectively access the structured data mainly published in RDF format, one needs to understand not only the prevalent vocabularies being used by the community, but also the extent and the patterns of its usage. In this paper, we achieve this by proposing a framework that analyzes the domain ontology usage and rank terms (classes, properties and attributes) based on multi-criteria characteristics that include population, coverage, and structure. We consider a purpose-built RDF dataset to select the popular terms and construct the schema based on the ranking, enabling the semantic web application to acquire information from the web of data effectively and efficiently.

Linked Open University Data applies semantic web and linked data technology to university data scenario, aiming at building interlinked semantic data around university information, providing possibility for unified inner- and inter- school information query and comparison. This paper proposes a general process of building linked open university data, with procedures covering choosing datasets and vocabularies, collecting and processing data, building RDF and interlink, etc. Tsinghua University Open Data is used to demonstrate the process. Tsinghua University consist of 5 well-formed, interconnected datasets, with a number of interesting applications has been built on top of them. Finally, remarkable points about data collecting and processing is discussed.

A matchmaking system for finding renting houses is required as the housing problem becomes serious in China and many people resort to rent a house. A semantic approach based on abductive conjunctive query answering (CQA) in Description Logic ontologies is exploited to provide more matches for a request about renting houses. Moreover, a matchmaking system based on this approach is developed. This demo will guide users to find suitable renting houses using this matchmaking system and show the advantages of the system.

A modern business process management (BPM) operates using common tenants of an underlying Service Oriented Architecture (SOA) runtime infrastructure based on the Service Component Architecture (SCA) and supports the BPMN 2.0 OMG

1

standard. Semantically-enabling all BPM artifacts, from high-level design to deployment and the runtime model of a BPM application, promotes continuous process refinement, comprehensive impact analysis, and reuse to minimize process and service proliferation. A semantic database can manage semantically-enabled BPM ontologies and models, enable machine-driven inference to discover implicit relationships in the models, and perform pattern-matching queries to find associations.

This paper presents an ontology for BPM based upon BPMN 2.0, Service Component Architecture (SCA) and the Web Ontology Language (OWL 2) that can support a wide range of use cases for process analysis, governance, business intelligence and systems management. It has the potential to bring together stakeholders across an enterprise, for a truly agile, end-to-end enterprise architecture.

This Demonstration paper discusses ongoing work at Tom Sawyer Software in the area of advanced visualization and analysis of very large data sets, including RDF data. There is a growing imperative to explore large data sets as part of opportunity and threat analysis in areas such as national defense, financial risk analysis, market intelligence, and disease epidemiology. An increasing volume of this type of information is represented as RDF graphs. By visualizing and visually analyzing data, it is possible to see patterns, trends, and outliers in complex RDF graphs that would otherwise be difficult or even impossible to discover. Since RDF graphs are by nature difficult for humans to read, Tom Sawyer Software has been developing an innovative, graphical approach to defining the schema of RDF data, visualizing salient parts of the RDF graph, and integrating social network analysis into the visualization process to provide intuitive visual navigation, query, and understanding of information. This Demonstration paper discusses the underlying technology and its realization in sophisticated software for building advanced data visualization and analysis applications for making sense of large RDF graphs.

Ontologies have been widely used in advanced information systems. However, it has been a challenging issue in ontology engineering to efficiently revise ontologies as new information becomes available. A novel method of revising ontologies has been proposed recently by Wang et al. However, related algorithms have not been implemented yet. In this article we describe an implementation of these algorithms called OntoRevision and report some experimental results. Our system is a plug-in for revising general ontologies in Protégé and thus can be used by Protégé users to revise ontologies automatically.

Ontology debugging helps users to understand the unsatisfiability of a concept in an ontology by finding minimal unsatisfiability-preserving sub-ontologies (MUPS) of the ontology for the concept. Although existing approaches have shown good performance for some real life ontologies, they are still inefficient to handle ontologies that have many MUPS for an unsatisfiable concept. In this paper, we propose an efficient approach to debugging ontologies based on a set of patterns. Patterns provide general information to explain unsatisfiability but are not dependent on a specific ontology. In this approach, we make use of a set of heuristic strategies and construct a directed graph w.r.t. the hierarchies where the depth-first search strategy can be used to search paths. The experiments show that our approach has gained a significant improvement over the state of the art and can find considerable number of MUPS.