Alessandro Adamou

Assessing the educational linked data landscape

In this research note, we present a preliminary study of available web datasets related to education, providing an overview of this area and, more importantly, highlighting how such linked datasets form a globally addressable network of resources for education. As expected, a certain level of heterogeneity was found. We therefore also show how a minor integration effort can improve the global cohesion of such a network of educational web data.

The Open University Linked Data – data.open.ac.uk

The article reports on the evolution of data.open.ac.uk, the Linked Open Data platform of the Open University, from a research experiment to a data hub for the open content of the University. Entirely based on Semantic Web technologies (RDF and the Linked Data principles), data.open.ac.uk is used to curate, publish and access data about academic degree qualifications, courses, scholarly publications and open educational resources of the University. It exposes a SPARQL endpoint and several other services to support developers, including queries stored server-side and entity lookup using known identifers such as course codes and YouTube video IDs. The platform is now a key information service at the Open University, with several core systems and websites exploiting linked data through data.open.ac.uk. Through these applications, data.open.ac.uk is now fulfilling a key role in the overall data infrastructure of the university, and in establishing connections with other educational institutions and information providers.

Kali-ma: a semantic guide to browsing and accessing functionalities in plugin-based tools

It is typical of plugin-based platforms such as the Eclipse RCP to be extensible by addition of functionalities. Such systems can undoubtedly benefit from having a vast developer community contributing to their enrichment. However, the proliferation of functionalities plugged in a system can be detrimental to its usability and bring confusion and clutter, as users who lack prior knowledge of the available features and how to access them can be unable to spot those that suit their needs. We present Kali-ma, a tool that equips Eclipse-based ontology engineering platforms with a GUI that allows users to browse and access plugin functionalities. The Kali-ma interface is dynamically generated by matching installed plugins with ontologies that describe their capabilities in the Semantic Web. It can adapt to selected criteria for classifying tools, and its approach is portable across systems supporting other domains, such as software engineering and business process management.

Building listening experience Linked Data through crowd-sourcing and reuse of library data

The Listening Experience Database (LED) is a project that gathers documented evidence of listening to music across cultural and historical contexts. Its underlying information system relies on the principles and practices of Linked Data, including a knowledge base that is itself a linked dataset structured according to common vocabularies for media and bibliographies such as Bibo and the Music Ontology. The data management workflow fully supports crowd-sourced input and incorporates data reuse from various sources right from the point of data collection, including the British National Bibliography dataset. The LED system gives contributors and moderators the tools to enhance these data by modelling divisions of bibliographc entries, providing information beyond the BNB schema, or aligning with other datasets by simple data reconciliation procedures. Vocabularies for music genre and instrument classes are also crowd-sourced on top of a baseline taxonomy from the DBpedia dataset. The Web frontend uses the Drupal content management system to encapsulate listening experiences into digital objects that incorporate textual information with references to literary, musical and other classes of related entities.

The foundations of virtual ontology networks

The notion of ontology network is relatively recent and rooted in the field of knowledge engineering. It concerns those ontology networks assembled at design time and established by their authors. Nowadays, however, the need to deal with heterogeneous semantic data sources, such as Linked Data and reengineered resources, gives rise to new use cases: ontologies now also need to be selected and combined in ways that were unpredictable at design time. This process is non-trivial and error-prone: if improperly dealt with, it can lead to loss of expressivity when interpreting resources (e.g. RDF graphs) as ontologies. Previous attempts at formally representing ontology relations assumed a controlled environment where axioms are unequivocally determined, and as such did not need to capture this distinction and its creeping issues. One possible solution is to assemble ontology networks at runtime by mimicking part of the design-time process; however, this requires that the difference between ontology networks assembled statically by their authors, and dynamically by consumer agents, be formally represented. The notion of virtual ontology network introduced here establishes this distinction. In this paper, we provide the theoretical underpinnings of virtual ontology network management. We define the representational primitives and abstract relationships that virtual networks are constructed upon. To validate the theory and demonstrate the feasibility of virtualization in contemporary ontology management, we show that this formal framework is compatible with the existing theory of ontology relations. To that end, we illustrate which constructs from existing representation languages can be used to implement the relations that characterize a virtual ontology network. We show this both in OWL 2 and with ε-connections. This paper is a prelude to actual virtualization methods that we have devised and implemented: the finalization of their evaluation is underway and will be presented as a follow-up to the work presented herein.

The NeOn Ontology Models

Interoperability on multiple levels, concerning both the ontologies themselves and their engineering activities, is a key requirement for ontology networks to be efficient, with minimal redundancy and high reuse. This requirement has a strict binding for software tools that can support some interoperability levels, yet they can be hindered by a lack of shared models and vocabularies describing the resources to be handled, as well as the ways of handling them. Here, three examples of metalevel vocabularies are proposed, each covering at least one peculiar interoperability aspect: OMV for modeling the artifacts themselves, LIR for managing a multilingual layer on top of them, and C-ODO Light for modeling collaboration-supportive life cycle management tasks and processes. All of these models lend themselves to handling by dedicated software tools and are all being employed within NeOn products.

A knowledge pattern-based method for linked data analysis

We present a Linked Data analysis method which relies on knowledge patterns for constructing a logical architecture of the knowledge in a dataset. This can then be exploited to compare heterogeneous datasets, enhance interoperability between them and make implicit knowledge emerge.

Supporting virtual integration of Linked Data with just-in-time query recompilation

Virtual data integration takes place at query execution time and relies on transformations of the original query to many target endpoints, where the data reside. In systems that integrate many data sources, this means maintaining many mappings, queries and query templates, as well as possibly issuing separate queries for linking entities in the datasets and retrieving their data. We propose a practical approach to keeping such complexity under control, which manipulates the translation from one client query to many target queries. The method performs just-in-time recompilation of the client query into elements that are combined with a query template into the target queries for multiple sources. It was validated in a setting with a custom star-shaped query language as client API and SPARQL endpoints as sources. The approach has shown to reduce the number of target queries to issue and of query templates to maintain, using a number of compiler functions that scales with the complexity of the data source, with an overhead that may be neglected where the method is most effective.

SPARQL Query Recommendations by Example

In this demo paper, a SPARQL Query Recommendation Tool (called SQUIRE) based on query reformulation is presented. Based on three steps, Generalization, Specialization and Evaluation, SQUIRE implements the logic of reformulating a SPARQL query that is satisfiable w.r.t a source RDF dataset, into others that are satisfiable w.r.t a target RDF dataset. In contrast with existing approaches, SQUIRE aims at recommending queries whose reformulations: (i) reflect as much as possible the same intended meaning, structure, type of results and result size as the original query and (ii) do not require to have a mapping between the two datasets. Based on a set of criteria to measure the similarity between the initial query and the recommended ones, SQUIRE demonstrates the feasibility of the underlying query reformulation process, ranks appropriately the recommended queries, and offers a valuable support for query recommendations over an unknown and unmapped target RDF dataset, not only assisting the user in learning the data model and content of an RDF dataset, but also supporting its use without requiring the user to have intrinsic knowledge of the data.

Software architectures for scalable ontology networks

Theory and practice in ontology management are drifting away from monolithic ontologies towards ontology networks. Processing interconnected knowledge models, e.g. through reasoning, can provide greater amounts of explicit knowledge, but at a high computational cost if the whole knowledge base has to be handled by concurrent processes. Our research tackles this problem via a software engineering approach. We devised a framework that combines privileged containers for ontology models with volatile containers for instance data that vary frequently. A reference implementation was developed in an Apache project for content management, and its adoption is providing data and use cases for validating it with objects from Fedora Commons repositories.