Gaia Varese

Ontology and instance matching

The growing need of sharing data and digital resources within and across organizations has produced a novel attention on issues related to ontology and instance matching. After an introductory classification of the main techniques and tools for ontology matching, the chapter focuses on instance matching by providing an accurate classification of the matching techniques proposed in the literature, and a comparison of the recent instance matching tools according to the results achieved in the OAEI 2009 contest. Ontology and instance matching solutions developed in the BOEMIE project for multimedia resource management and ontology evolution are finally presented.

Semantic coordination of P2P collective intelligence

P2P techniques for semantic coordination based on semantic communities are recently emerging to enforce effective collaboration platforms. In this paper, we focus on semantic coordination in the iCoord P2P system and we introduce the notion of collective intelligence to enable semantic peer communities to setup a shared community vocabulary which is gradually built with the autonomous contribution of all the community members according to their terminological preferences. Techniques for extending iCoord to include in the collective intelligence also the social peer knowledge coming from conventional tagging systems are also presented.

Ontology Knowledge Authoring by Natural Language Empowerment

One of the main limitations in using ontologies in modern information systems is due to the fact that the design and maintenance of ontologies is a complex activity that requires a specific background in formal logics and data semantics. In this paper, we present the iCoord coordination system, which aims at supporting content managers and domain experts in the process of defining and maintaining domain ontologies. In particular, we describe the Natural Language Explanation tool (NLE) of iCoord for ontology knowledge authoring.

Matching Semantic Web Resources

In this paper, we propose knowledge-chunk based techniques for instance matching and mapping discovery of Semantic Web resources. Knowledge chunks provide a synthetic representation of semantic descriptions of individuals with the aim to improve matching efficiency also in presence of rich and articulated RDF/OWL resource descriptions for the domain. The main implementation choices in the HMatch 2.0 system and the evaluation results obtained on a corpus of web resource descriptions in the Athletics domain are also presented.

Similarity-Based Classification of Microdata

In this paper, we propose a similarity-based approach for microdata classification based on tagging, matching and clouding techniques. The goal is to construct entity-centric microdata clouds where similar microdata items can be properly arranged to highlight their relevance with respect to a selected target entity according to different notions of relevance defined in the paper. An application example is provided, based on a microdata collection extracted from a real microblogging system.

Building Collective Tag Intelligence through Folksonomy Coordination

In this chapter, we provide techniques for automatically classifying and coordinating tags extracted from one or more folksonomies, with the aim of building collective tag intelligence which can then be exploited to improve the conventional searching functionalities provided by tagging systems. Collective tag intelligence is organized in form of tag equivalence clusters with corresponding semantic, terminological, and linguistic relations. For building tag collective intelligence, we define i) normalization techniques to identify equivalence clusters of tags and extract the relations holding between them and ii) similarity techniques to match tags on the basis of available collective tag intelligence. Finally, we describe the evaluation of the proposed techniques over real datasets extracted from del.icio.us and Flickr folksonomies and a real application example of exploiting the collective tag intelligence for similarity-based resource retrieval.

Managing Collective Intelligence in Semantic Communities of Interest

In this paper, the authors present a reference P2P architecture based on autonomous, self-emerging semantic communities of interest CoIs for collective intelligence creation and management. An approach for enabling knowledge organization and management at the level of a single peer is presented in the paper, as well as techniques for supporting a peer to participate to the construction of a shared community vocabulary, according to the terminological preferences automatically extracted from its personal knowledge. Furthermore, an application example in the e-health domain is presented in the framework of the iCoord system for P2P semantic coordination to show the use of a manifesto-based collective intelligence for enforcing effective collaboration in a real case study.