In this paper, we study query evaluation on Active XML documents (AXML for short), a new generation of XML documents that has recently gained popularity. AXML documents are XML documents whose content is given partly extensionally, by explicit data elements, and partly intensionally, by embedded calls to Web services, which can be invoked to generate data.A major challenge in the efficient evaluation of queries over such documents is to detect which calls may bring data that is relevant for the query execution, and to avoid the materialization of irrelevant information. The problem is intricate, as service calls may be embedded anywhere in the document, and service invocations possibly return data containing calls to new services. Hence, the detection of relevant calls becomes a continuous process. Also, a good analysis must take the service signatures into consideration.We formalize the problem, and provide algorithms to solve it. We also present an implementation that is compliant with XML and Web services standards, and is used as part of the ActiveXML system. Finally, we experimentally measure the performance gains obtained by a careful filtering of the service calls to be triggered.

No abstract

No abstract

We present KADOP, a

The Semantic Web has made huge progress in the last decade, and now comprises hundreds of knowledge bases (KBs). The Linked Open Data cloud connects the KBs in this Web of data. However, the links between the KBs are mostly concerned with the instances, not with the schema. Aligning the schemas is not easy, because the KBs may differ not just in their names for relations and classes, but also in their inherent structure. Therefore, we argue in this paper that advanced schema alignment is needed to tie the Semantic Web together. We put forward a particularly simple approach to illustrate how that might look.

Many data providers make their data available through Web service APIs. In order to unleash the potential of these sources for intelligent applications, the data has to be combined across different APIs. However, due to the heterogeneity of schemas, the integration of different APIs remains a mainly manual task to date. In this paper, we model an API method as a view with binding patterns over a global RDF schema. We present an algorithm that can automatically infer the view definition of a method in the global schema. We also show how to compute transformation functions that can transform API call results into this schema. The key idea of our approach is to exploit the intersection of API call results with a knowledge base and with other call results. Our experiments on more than 50 real Web services show that we can automatically infer the schema with a precision of 81%-100%.

No abstract

The large number of linked datasets in the Web, and their diversity in terms of schema representation has led to a fragmented dataset landscape. Querying and addressing information needs that span across disparate datasets requires the alignment of such schemas. Majority of schema and ontology alignment approaches focus exclusively on class alignment. Yet, relation alignment has not been fully addressed, and existing approaches fall short on addressing the dynamics of datasets and their size. In this work, we address the problem of relation alignment across disparate linked datasets. Our approach focuses on two main aspects. First, online relation alignment, where we do not require full access, and sample instead for a minimal subset of the data. Thus, we address the main limitation of existing work on dealing with the large scale of linked datasets, and in cases where the datasets provide only query access. Second, we learn supervised machine learning models for which we employ various features or matchers that account for the diversity of linked datasets at the instance level. We perform an experimental evaluation on real-world linked datasets, DBpedia, YAGO, and Freebase. The results show superior performance against state-of-the-art approaches in schema matching, with an average relation alignment accuracy of 84%. In addition, we show that relation alignment can be performed efficiently at scale.