Mathieu d'Aquin

Toward a New Generation of Semantic Web Applications

Although research on integrating semantics with the Web started almost as soon as the Web was in place, a concrete Semantic Web that is, a large-scale collection of distributed semantic metadata emerged only over the past four to five years. The Semantic Webs embryonic nature is reflected in its existing applications. Most of these applications tend to produce and consume their own data, much like traditional knowledge- based applications, rather than actually exploiting the Semantic Web as a large-scale information source. These first-generation semantic Web applications typically use a single ontology that supports integration of resources selected at design time.

Exploring the Semantic Web as Background Knowledge for Ontology Matching

In this paper we propose an ontology matching paradigm based on the idea of harvesting the Semantic Web, i.e., automatically finding and exploring multiple and heterogeneous online knowledge sources to derive mappings. We adopt an experimental approach in the context of matching two real life, large-scale ontologies to investigate the potential of this paradigm, its limitations, and its relation to other techniques. Our experiments yielded a promising baseline precision of 70% and identified a set of critical issues that need to be considered to achieve the full potential of the paradigm. Besides providing a good performance as a stand-alone matcher, our paradigm is complementary to existing techniques and therefore could be used in hybrid tools that would further advance the state of the art in the ontology matching field.

Identifying Key Concepts in an Ontology, through the Integration of Cognitive Principles with Statistical and Topological Measures

In this paper we address the issue of identifying the concepts in an ontology, which best summarize what the ontology is about. Our approach combines a number of criteria, drawn from cognitive science, network topology, and lexical statistics. In the paper we show two versions of our algorithm, which have been evaluated against the results produced by human experts. We report that the latest version of the algorithm performs very well, exhibiting an excellent degree of correlation with the choices of the experts. While the generation of automatic methods for ontology summarization is an interesting research issue in itself, the work described here also provides a basis for novel approaches to a variety of ontology engineering tasks, including ontology matching, automatic classification, ontology modularization, and ontology evaluation.

Ontology evolution: a process-centric survey

Ontology evolution aims at maintaining an ontology up to date with respect to changes in the domain that it models or novel requirements of information systems that it enables. The recent industrial adoption of Semantic Web techniques, which rely on ontologies, has led to the increased importance of the ontology evolution research. Typical approaches to ontology evolution are designed as multiple-stage processes combining techniques from a variety of fields (e.g., natural language processing and reasoning). However, the few existing surveys on this topic lack an in-depth analysis of the various stages of the ontology evolution process. This survey extends the literature by adopting a process-centric view of ontology evolution. Accordingly, we first provide an overall process model synthesized from an overview of the existing models in the literature. Then we survey the major approaches to each of the steps in this process and conclude on future challenges for techniques aiming to solve that particular stage.

Unsupervised learning of link discovery configuration

Discovering links between overlapping datasets on the Web is generally realised through the use of fuzzy similarity measures. Configuring such measures is often a non-trivial task that depends on the domain, ontological schemas, and formatting conventions in data. Existing solutions either rely on the users knowledge of the data and the domain or on the use of machine learning to discover these parameters based on training data. In this paper, we present a novel approach to tackle the issue of data linking which relies on the unsupervised discovery of the required similarity parameters. Instead of using labeled data, the method takes into account several desired properties which the distribution of output similarity values should satisfy. The method includes these features into a fitness criterion used in a genetic algorithm to establish similarity parameters that maximise the quality of the resulting linkset according to the considered properties. We show in experiments using benchmarks as well as real-world datasets that such an unsupervised method can reach the same levels of performance as manually engineered methods, and how the different parameters of the genetic algorithm and the fitness criterion affect the results for different datasets.

Watson, more than a Semantic Web search engine

In this tool report, we present an overview of the Watson system, a Semantic Web search engine providing various functionalities not only to find and locate ontologies and semantic data online, but also to explore the content of these semantic documents. Beyond the simple facade of a search engine for the Semantic Web, we show that the availability of such a component brings new possibilities in terms of developing semantic applications that exploit the content of the Semantic Web. Indeed, Watson provides a set of APIs containing high level functions for finding, exploring and querying semantic data and ontologies that have been published online. Thanks to these APIs, new applications have emerged that connect activities such as ontology construction, matching, sense disambiguation and question answering to the Semantic Web, developed by our group and others. In addition, we also describe Watson as a unprecedented research platform for the study the Semantic Web, and of formalised knowledge in general.

Ontology modularization for knowledge selection: experiments and evaluations

Problems with large monolithical ontologies in terms of reusability, scalability and maintenance have led to an increasing interest in modularization techniques for ontologies. Currently, existing work suffers from the fact that the notion of modularization is not as well understood in the context of ontologies as it is in software engineering. In this paper, we experiment on applying state-of-the-art tools for ontology modularization in the context of a concrete application: the automatic selection of knowledge components to be used for Web page annotation and semantic browsing. We conclude that, in a broader context, an evaluation framework is required to guide the choice of a modularization tool, in accordance with the requirements of the considered application.

ADAPTATION KNOWLEDGE ACQUISITION: A CASE STUDY FOR CASE-BASED DECISION SUPPORT IN ONCOLOGY

Kasimir is a case‐based decision support system in the domain of breast cancer treatment. For this system, a problem is given by the description of a patient and a solution is a set of therapeutic decisions. Given a target problem, Kasimir provides several suggestions of solutions, based on several justified adaptations of source cases. Such adaptation processes are based on adaptation knowledge. The acquisition of this kind of knowledge from experts is presented in this paper. It is shown how the decomposition of adaptation processes by introduction of intermediate problems can highlight simple and generalizable adaptation steps. Moreover, some adaptation knowledge units that are generalized from those acquired for Kasimir are presented. This knowledge can be instantiated in other case‐based decision support systems, in particular in medicine.

A novel approach to visualizing and navigating ontologies

There is empirical evidence that the user interaction metaphors used in ontology engineering toolkits are largely inadequate and that novel interactive frameworks for human ontology interaction are needed. Here we present a novel tool for visualizing and navigating ontologies, called KC Viz, which exploits an innovative ontology summarization method to support a ’middleout ontology browsing’ approach, where it becomes possible to navigate ontologies starting from the most information-rich nodes (i.e., key concepts). This approach is similar to map-based visualization and navigation in Geographical Information Systems, where, e.g., major cities are displayed more prominently than others, depending on the current level of granularity.

Criteria and Evaluation for Ontology Modularization Techniques

While many authors have argued for the benefits of applying principles of modularization to ontologies, there is not yet a common understanding of how modules are defined and what properties they should have. In the previous section, this question was addressed from a purely logical point of view. In this chapter, we take a broader view on possible criteria that can be used to determine the quality of a modules. Such criteria include logic-based, but also structural and application-dependent criteria, sometimes borrowing from related fields such as software engineering. We give an overview of possible criteria and identify a lack of application-dependent quality measures. We further report some modularization experiments and discuss the role of quality criteria and evaluation in the context of these experiments.