Search

The emergence of open information extraction as a tool for constructing and expanding knowledge graphs has aided the growth of temporal data, for instance, YAGO, NELL and Wikidata. While YAGO and Wikidata maintain the valid time of facts, NELL records the time point at which a fact is retrieved from some Web corpora. Collectively, these knowledge graphs (KG) store facts extracted from Wikipedia and other sources. Due to the imprecise nature of the extraction tools that are used to build and expand KG, such as NELL, the facts in the KG are weighted (a confidence value representing the correctness of a fact). Additionally, NELL can be considered as a transaction time KG because every fact is associated with extraction date. On the other hand, YAGO and Wikidata use the valid time model because they maintain facts together with their validity time (temporal scope). In this paper, we propose a bitemporal model (that combines transaction and valid time models) for maintaining and querying bitemporal probabilistic knowledge graphs. We study coalescing and scalability of marginal and MAP inference. Moreover, we show that complexity of reasoning tasks in atemporal probabilistic KG carry over to the bitemporal setting. Finally, we report our evaluation results of the proposed model.

Argumentative relation classification is the task of determining the type of relation (e.g., support or attack) that holds between two argument units. Current state-of-the-art models primarily exploit surface-linguistic features including discourse markers, modals or adverbials to classify argumentative relations. However, a system that performs argument analysis using mainly rhetorical features can be easily fooled by the stylistic presentation of the argument as opposed to its content, in cases where a weak argument is concealed by strong rhetorical means. This paper explores the difficulties and the potential effectiveness of knowledge-enhanced argument analysis, with the aim of advancing the state-of-the-art in argument analysis towards a deeper, knowledge-based understanding and representation of arguments. We propose an argumentative relation classification system that employs linguistic as well as knowledge-based features, and investigate the effects of injecting background knowledge into a neural baseline model for argumentative relation classification. Starting from a Siamese neural network that classifies pairs of argument units into support vs. attack relations, we extend this system with a set of features that encode a variety of features extracted from two complementary background knowledge resources: ConceptNet and DBpedia. We evaluate our systems on three different datasets and show that the inclusion of background knowledge can improve the classification performance by considerable margins. Thus, our work offers a first step towards effective, knowledge-rich argument analysis.

Time-wise knowledge is relevant in knowledge graphs as the majority facts are true in some time period, for instance, (Barack Obama, president of, USA, 2009, 2017). Consequently, temporal information extraction and temporal scoping of facts in knowledge graphs have been a focus of recent research. Due to this, a number of temporal knowledge graphs have become available such as YAGO and Wikidata. In addition, since the temporal facts are obtained from open text, they can be weighted, i.e., the extraction tools assign each fact with a confidence score indicating how likely that fact is to be true. Temporal facts coupled with confidence scores result in a probabilistic temporal knowledge graph. In such a graph, probabilistic query evaluation (marginal inference) and computing most probable explanations (MPE inference) are fundamental problems. In addition, in these problems temporal coalescing, an important research in temporal databases, is very challenging. In this work, we study these problems by using probabilistic programming. We report experimental results comparing the efficiency of several state of the art systems.

Stream Data processing has become a popular topic in database research addressing the challenge of efficiently answering queries over continuous data streams. Meanwhile data streams have become more and more important as a basis for higher level decision processes that require complex reasoning over data streams and rich background knowledge. In previous work the foundation for complex reasoning over streams and background knowledge was laid by introducing technologies for wrapping and querying streams in the RDF data format and by supporting simple forms of reasoning in terms of incremental view maintenance. In this paper, we discuss how this existing technologies should be extended toward richer forms of reasoning using Sensor Networks as a motivating example.

Several countries are currently working on setting up geoportals as part of their national spatial data infrastructure (SDI) (and this is also a requirement of the Inspire initiative). A key ability of these geoportals is that the user should be able to view (and download) data from several sources from one access point. This will certainly make the access to geospatial data easier. However, there are also cartographic and semantic challenges that have to be solved. In this discussion group we discussed some topics concerning both download services and view services and some possible solutions.

The goal of semantic web research is to turn the World-Wide Web into a Web of Data that can be processed automatically to a much larger extend than possible with traditional web technology. Important features of the solution currently being developed is the ability to link data from from different sources and to provide formal definitions of the intended meaning of the terminology used in different sources as a basis for deriving implicit information and for conflict detection. Both requires the ability to reason about the definition of terms. With the development of OWL as the standard language for representing terminological knowledge, reasoning in description logics has been determined as the major technique for performing this reasoning cite{OWLreasoning}. More recently, rule languages have gained more importance as well as they have been shown to be more suited for efficient reasoning about terminology and data at the same time. So far little attention has been paid to the problem of representing and reasoning about space and time on the semantic web. In particular, existing semantic web languages are not well suited for representing these aspects as they require to operate over metric spaces that behave fundamentally different from the abstract interpretation domains description logics are based on. Nevertheless, there is a strong need to integrate reasoning about space and time into existing semantic web technologies especially because more and more data available on the web has a references to space and time. Images taken by digital cameras are a good example of such data as they come with a time stamp and geographic coordinates. In this paper, we concentrate on spatial aspects and discuss different use case for reasoning about spatial aspects on the (semantic) web and possible technological solutions for these use cases. Based on these discussions we conclude that the actual open problem is not existing technologies for terminological or spatial reasoning

In many scientific domains, researchers depend on a timely and efficient access to available publications in their particular area. The increasing availability of publications in electronic form via digital libraries is a reaction to this need. A remaining problem is the fact that the pool of all available publications is distributed between different libraries. In order to increase the availability of information, these different libraries should be linked in such a way, that all the information is available via any one of them. Peer-to-peer technologies provide sophisticated solutions for this kind of loose integration of information sources. In our work, we consider digital libraries that organize documents according to a dedicated classification hierarchy or provide access to information on the basis of a thesaurus. These kinds of access mechanisms have proven to increase the retrieval result and are therefore widely used. On the other hand, this causes new problems as different sources will use different classifications and thesauri to organize information. This means, that we have to be able to mediate between these different structures. Integrating this mediation into the information retrieval process is a problem that to the best of our knowledge has not been addressed before.

The aim of this breakout session was to chart the landscape of existing approaches for representing mappings between heterogeneous models, identify common ideas and formulate research questions to be addressed in the future. In the session, the discussion mainly concerned three aspects: The nature of mappings, existing proposals for mappings and open research questions.

Ontology mapping, alignment, and translation has been an active research component of the general research on semantic integration and interoperability. In our talk, we gave our own classification of different topics in this research. We talked about types of heterogeneity between ontologies, various mapping representations, classified methods for discovering methods both between ontology concepts and data, and talked about various tasks where mappings are used. In this extended abstract of our talk, we provide an annotated bibliography for this area of research, giving readers brief pointers on representative papers in each of the topics mentioned above. We did not attempt to compile a comprehensive bibliography and hence the list in this abstract is necessarily incomplete. Rather, we tried to sketch a map of the field, with some specific reference to help interested readers in their exploration of the work to-date.