Interacting with knowledge graphs can be a daunting task for people without a background in computer science since the query language that is used (SPARQL) has a high barrier of entry. Large language models (LLMs) can lower that barrier by providing support in the form of Text2SPARQL translation. In this paper we introduce a generalized method based on SPINACH, an LLM backed agent that translates natural language questions to SPARQL queries not in a single shot, but as an iterative process of exploration and execution. We describe the overall architecture and reasoning behind our design decisions, and also conduct a thorough analysis of the agent behavior to gain insights into future areas for targeted improvements. This work was motivated by the Text2SPARQL challenge, a challenge that was held to facilitate improvements in the Text2SPARQL domain.

This is the report of our participation in the KGCW2023 Challenge @ ESWC 2023 with our RDFProcessing-Toolkit/Sansa system which won the “fastest” tool award. The challenge was about the construction of RDF knowledge graphs from RML specifications with varying complexity in regard to the mix of input formats, characteristics of the data and the needed join operations. We detail how we integrated our tool into the provided benchmark framework. Thereby we also report on the issues and shortcomings we encountered as a base for future improvements. Furthermore, we provide an analysis of the data measured with the benchmark framework.

Approaches for the construction of knowledge graphs from heterogeneous data sources range from ad-hoc scripts to dedicated mapping languages. Two common foundations are thereby RML and SPARQL. So far, both approaches are treated as different: On the one hand there are tools specifically for processing RML whereas on the other hand there are tools that extend SPARQL in order to incorporate additional data sources. In this work, we first show how this gap can be bridged by translating RML to a sequence of SPARQL CONSTRUCT queries and introduce the necessary SPARQL extensions. In a subsequent step, we employ techniques to optimize SPARQL query workloads as well as individual query execution times in order to obtain an optimized sequence of queries with respect to the order and uniqueness of the generated triples. Finally, we present a corresponding SPARQL query execution engine based on the Apache Spark Big Data framework. In our evaluation on benchmarks we show that our approach is capable of achieving RML mapping execution performance that surpasses the current state of the art.

In the context of crisis informatics, the integration and exploitation of high volumes of heterogeneous data from multiple sources is one of the big chances as well as challenges up to now. Semantic Web technologies have proven a valuable means to integrate and represent knowledge on the basis of domain concepts which improves interoperability and interpretability of information resources and allows deriving more knowledge via semantic relations and reasoning. In this paper, we investigate the potential of representing and processing geospatial information within the semantic paradigm. We show, on the technical level, how existing open source means can be used and supplemented as to efficiently handle geographic information and to convey exemplary results highly relevant in context of crisis management applications. When given semantic resources get enriched with geospatial information, new information can be retrieved combining the concepts of multi-polygons and geo-coordinates and using the power of GeoSPARQL queries. Custom SPARQL extension functions and data types for JSON, XML and CSV as well as for dialects such as GeoJSON and GML allow for succinct integration of heterogeneous data. We implemented these features for the Apache Jena Semantic Web framework by leveraging its plugin systems. Furthermore, significant improvements w.r.t. GeoSPARQL query performance have been contributed to the framework.

Geospatial knowledge has always been an essential driver for many societal aspects. This concerns in particular urban planning and urban growth management. To gain insights from geospatial data and guide decisions usually authoritative and open data sources are used, combined with user or citizen sensing data. However, we see a great potential for improving geospatial analytics by combining geospatial data with the rich terminological knowledge, e.g., provided by the Linked Open Data Cloud. Having semantically explicit, integrated geospatial and terminological knowledge, expressed by means of established vocabularies and ontologies, cross-domain spatial analytics can be performed. One analytics technique working on terminological knowledge is inductive concept learning, an approach that learns classifiers expressed as logical concept descriptions. In this paper, we extend inductive concept learning to infer and make use of the spatial context of entities in spatio-terminological data. We propose a formalism for extracting and making spatial relations explicit such that they can be exploited to learn spatial concept descriptions, enabling 'spatially aware' concept learning. We further provide an implementation of this formalism and demonstrate its capabilities in different evaluation scenarios.

In this work, we present a schema-agnostic faceted browsing benchmark generation framework for RDF data and SPARQL engines. Faceted search is a technique that allows narrowing down sets of information items by applying constraints over their properties, whereas facets correspond to properties of these items. While our work can be used to realise real-world faceted search user interfaces, our focus lies on the construction and benchmarking of faceted search queries over knowledge graphs. The RDF model exhibits several traits that seemingly make it a natural foundation for faceted search: all information items are represented as RDF resources, property values typically already correspond to meaningful semantic classifications, and with SPARQL there is a standard language for uniformly querying instance and schema information. However, although faceted search is ubiquitous today, it is typically not performed on the RDF model directly. Two major sources of concern are the complexity of query generation and the query performance. To overcome the former, our framework comes with an intermediate domain-specific language. Thereby our approach is SPARQL-driven which means that every faceted search information need is intensionally expressed as a single SPARQL query. In regard to the latter, we investigate the possibilities and limits of real-time SPARQL-driven faceted search on contemporary triple stores. We report on our findings by evaluating systems performance and correctness characteristics when executing a benchmark generated using our generation framework. All components, namely the benchmark generator, the benchmark runners and the underlying faceted search framework, are published freely available as open source. (C) 2020 Elsevier B.V. All rights reserved.

With the ever-growing generation of data for the Semantic Web comes an increasing demand for this data to be made available to non-semantic Web experts. One way of achieving this goal is to translate the languages of the Semantic Web into natural language. We present LD2NL, a framework for verbalizing the three key languages of the Semantic Web, i.e., RDF, OWL, and SPARQL. Our framework is based on a bottom-up approach to verbalization. We evaluated LD2NL in an open survey with 86 persons. Our results suggest that our framework can generate verbalizations that are close to natural languages and that can be easily understood by non-experts. Therewith, it enables non-domain experts to interpret Semantic Web data with more than 91% of the accuracy of domain experts.

Automatic subject indexing has been a longstanding goal of digital curators to facilitate effective retrieval access to large collections of both online and offline information resources. Controlled vocabularies are often used for this purpose, as they standardise annotation practices and help users to navigate online resources through following interlinked topical concepts. However, to this date, the assignment of suitable text annotations from a controlled vocabulary is still largely done manually, or at most (semi-)automatically, even though effective machine learning tools are already in place. This is because existing procedures require a sufficient amount of training data and they have to be adapted to each vocabulary, language and application domain anew. In this paper, we argue that there is a third solution to subject indexing which harnesses cross-domain knowledge graphs. Our KINDEX approach fuses distributed knowledge graph information from different sources. Experimental evaluation shows that the approach achieves good accuracy scores by exploiting correspondence links of publicly available knowledge graphs.

The following paper is an extended summary of the journal paper "DL-Learner A framework for inductive learning on the Semantic Web". In this system paper, we describe the DL-Learner framework. It is beneficial in various data and schema analytic tasks with applications in different standard machine learning scenarios, e.g. life sciences, as well as Semantic Web specific applications such as ontology learning and enrichment. Since its creation in 2007, it has become the main OWL and RDF-based software framework for supervised structured machine learning and includes several algorithm implementations, usage examples and has applications building on top of the framework.

The availability of structured data has increased significantly over the past decade and several approaches to learn from structured data have been proposed. These logic-based, inductive learning methods are often conceptually similar, which would allow a comparison among them even if they stem from different research communities. However, so far no efforts were made to define an environment for running learning tasks on a variety of tools, covering multiple knowledge representation languages. With SML-Bench, we propose a benchmarking framework to run inductive learning tools from the ILP and semantic web communities on a selection of learning problems. In this paper, we present the foundations of SML-Bench, discuss the systematic selection of benchmarking datasets and learning problems, and showcase an actual benchmark run on the currently supported tools.

We demonstrate the open-source Semantic Analytics Stack (SANSA), which can perform scalable analysis of large-scale knowledge graphs to facilitate applications such as link prediction, knowledge base completion and reasoning. The motivation behind this work lies in the lack of scalable methods for analytics which exploit expressive structures underlying semantically structured knowledge bases. The demonstration is based on the BigDataEurope technical platform, which utilizes Docker technology. We present various examples of using SANSA in the form of interactive Spark notebooks, which are executed with Apache Zeppelin. The technical platform and the notebooks are available on SANSA Github and can be deployed on any Docker-enabled host, locally or in a Docker Swarm cluster.

A major research challenge is to perform scalable analysis of large-scale knowledge graphs to facilitate applications like link prediction, knowledge base completion and reasoning. Analytics methods which exploit expressive structures usually do not scale well to very large knowledge bases, and most analytics approaches which do scale horizontally (i.e., can be executed in a distributed environment) work on simple feature-vector-based input. This software framework paper describes the ongoing Semantic Analytics Stack (SANSA) project, which supports expressive and scalable semantic analytics by providing functionality for distributed computing on RDF data.

The problem of predicting the membership w.r.t. a target concept for individuals of Semantic Web knowledge bases can be cast as a concept learning problem, whose goal is to induce intensional definitions describing the available examples. However, the models obtained through the methods borrowed from Inductive Logic Programming e.g. Terminological Decision Trees, may be affected by two crucial aspects: the refinement operators for specializing the concept description to be learned and the heuristics employed for selecting the most promising solution (i.e. the concept description that describes better the examples). In this paper, we started to investigate the effectiveness of Terminological Decision Tree and its evidential version when a refinement operator available in DL-Learner and modified heuristics are employed. The evaluation showed an improvement in terms of the predictiveness.

We show how to convert OWL Class Expressions to SPARQL queries where the instances of that concept are with a specific ABox equal to the SPARQL query result. Furthermore, we implement and integrate our converter into the CELOE algorithm (Class Expression Learning for Ontology Engineering), where it replaces the position of a traditional OWL reasoner. This will foster the application of structured machine learning to the Semantic Web, since most data is readily available in triple stores. We provide experimental evidence for the usefulness of the bridge. In particular, we show that we can improve the run time of machine learning approaches by several orders of magnitude.

In this system paper, we describe the DL-Learner framework, which supports supervised machine learning using OWL and RDF for background knowledge representation. It can be beneficial in various data and schema analysis tasks with applications in different standard machine learning scenarios, e.g. in the life sciences, as well as Semantic Web specific applications such as ontology learning and enrichment. Since its creation in 2007, it has become the main OWL and RDF-based software framework for supervised structured machine learning and includes several algorithm implementations, usage examples and has applications building on top of the framework. The article gives an overview of the framework with a focus on algorithms and use cases.

The Linked Open Data Cloud is a goldmine for creating open and low-cost educational applications: First, it contains open knowledge of encyclopedic nature on a large number of real-world entities. Moreover, the data being structured ensures that the data is both human- and machine-readable. Finally, the openness of the data and the use of RDF as standard format facilitate the development of applications that can be ported across different domains with ease. However, RDF is still unknown to most members of the target audience of educational applications. Thus, Linked Data has commonly been used for the description or annotation of educational data. Yet, Linked Data has (to the best of our knowledge) never been used as direct source of educational material. With ASSESS, we demonstrate that Linked Data can be used as a source for the automatic generation of educational material. By using innovative RDF verbalization and entity summarization technology, we bridge between natural language and RDF. We then use RDF data directly to generate quizzes which encompass questions of different types on user-defined domains of interest. By these means, we enable learners to generate self-assessment tests on domains of interest. Our evaluation shows that ASSESS generates high-quality English questions. Moreover, our usability evaluation suggests that our interface can be used intuitively. Finally, our test on DBpedia shows that our approach can be deployed on very large knowledge bases.

The publishing industry is at the verge of an era, wherein particular professional customers of publishing products are not so much interested in comprehensive books and journals, i.e. traditional publishing products, anymore as they now are interested in possibly structured information pieces delivered just-in-time as a certain information need arises. This requires a transformation of the publishing workflows towards the production of much richer meta-data for fine-grained and highly interlinked pieces of content. Linked Data can play a crucial role in this transition. The LOD2 Stack is an integrated distribution of aligned tools which support the whole lifecycle of Linked Data from extraction, authoring/creation via enrichment, interlinking, fusing to maintenance. In this application paper, we describe a real-world usage scenario of the LOD2 stack at a global publishing company. We give an overview over the LOD2 Stack and the underlying life-cycle of Linked Data, describe data-flows and usage scenarios at a publisher and then show how the stack supports those scenarios.

The decentral architecture behind theWeb has led to pieces of information being distributed across data sources with varying structure. Hence, answering complex questions often requires combining information from structured and unstructured data sources. We present HAWK, a novel entity search approach for Hybrid Question Answering based on combining Linked Data and textual data. The approach uses predicate-argument representations of questions to derive equivalent combinations of SPARQL query fragments and text queries. These are executed so as to integrate the results of the text queries into SPARQL and thus generate a formal interpretation of the query. We present a thorough evaluation of the framework, including an analysis of the influence of entity annotation tools on the generation process of the hybrid queries and a study of the overall accuracy of the system. Our results show that HAWK achieves 0.68 respectively 0.61 F-measure within the training respectively test phases on the Question Answering over Linked Data (QALD-4) hybrid query benchmark.

One of the main tasks when creating and maintaining knowledge bases is to validate facts and provide sources for them in order to ensure correctness and traceability of the provided knowledge. So far, this task is often addressed by human curators in a three-step process: issuing appropriate keyword queries for the statement to check using standard search engines, retrieving potentially relevant documents and screening those documents for relevant content. The drawbacks of this process are manifold. Most importantly, it is very time-consuming as the experts have to carry out several search processes and must often read several documents. In this article, we present DeFacto (Deep Fact Validation)-an algorithm able to validate facts by finding trustworthy sources for them on the Web. DeFacto aims to provide an effective way of validating facts by supplying the user with relevant excerpts of web pages as well as useful additional information including a score for the confidence DeFacto has in the correctness of the input fact. To achieve this goal, DeFacto collects and combines evidence from web pages written in several languages. In addition, DeFacto provides support for facts with a temporal scope, i.e., it can estimate in which time frame a fact was valid. Given that the automatic evaluation of facts has not been paid much attention to so far, generic benchmarks for evaluating these frameworks were not previously available. We thus also present a generic evaluation framework for fact checking and make it publicly available.