The application of semantic web technologies to multimedia data fusion within escience

Abstract

Advances in scientific research techniques have led to an explosion of informationrich, multimedia data within the research sector. New high-throughput data capture and combinatorial experimentation techniques (involving advanced instruments capable of capturing extremely high-resolution data streams) have resulted in the generation of research data in quantities that are too great for effective assimilation. The data is not only massive in volume but is also being produced in a broad range of mediums and formats, including numerical data, spectrographic output, genomic arrays, images, 3D models, audio and video, for disciplines including nano-materials, bioinformatics, tele-medicine, geosciences, astronomy and the social sciences. Scientific discovery is increasingly dependent on reliable tools and services to support the storage, dissemination, analysis and correlation of these complex data sets by collaborating teams of globally distributed scientists. The volume, variety and multi-dimensional nature of the content exacerbate the difficulty of describing this data adequately so it can be confidently and appropriately incorporated into existing theories or models. In order to validate and authenticate scientific results, detailed provenance metadata describing the precise methodology and derived data sets needs to be recorded. Because todays scientists are working in large geographically distributed teams or "virtual organisations, the data and metadata have to be comprehensible to people, computers and software across many different organisations, platforms and disciplines. Metadata standards and semantic interoperability are essential to enable distributed querying, analysis and integration of mixed-media and heterogeneous scientific data sets in order to maximise their reuse, extract the inherent knowledge and build new knowledge The Semantic Web (Berners-Lee, Hendler and Lassila 2001) promotes interoperability through formal languages and rich semantics. It aims to build a web where information is exchanged easily between humans and machines. Chapter 2 describes the layered standards and protocols for data definition, storage and exchange that make up the Semantic Web architecture: eXtensible Markup Language (XML) (Bray, Paoli, Sperberg-McQueen,Maler, Yergeau and Cowan 2006), Resource Description Framework (RDF) (Beckett 2004), Web Ontology Language (OWL) (McGuinness and van Harmelen 2004) and Uniform Resource Identifiers (URIs) (Berners-Lee, Fielding andMasinter 2005). Through a combination of these technologies, the SemanticWeb aims to define and expose the semantics associated with data or information, in order to facilitate automatic processing, integration, sharing and reuse of the data. Our hypothesis is that the application of Semantic Web technologies to the semantic annotation, integration and correlation of distributed mixed-media scientific data sets and scientific data processing services offers enormous potential for expediting the discovery of new knowledge. SemanticWeb/grid tools enhance interoperability through formal syntaxes, ontologies and inferencing rules. They enable innovative search, data exploration, hypothesis development and evaluation interfaces and can assist researchers in managing, assimilating and distributing data to facilitate further scientific understanding and discovery. The remainder of this chapter is structured as follows. In the next section (Section 8.2), we describe some of the key challenges and Semantic Web technologies either currently available or emerging that could usefully be applied to eResearch or eScience problems. In Sections 8.3-8.5, we describe three case studies in which we applied, evaluated and extended Semantic Web technologies: fuel cell optimisation semantic WildNet EthnographicMedia Analysis. Finally, in Section 8.6, we conclude with a brief discussion on the value-add of SemanticWeb/grid technologies and where SemanticWeb technologies are heading in the context of scientific multimedia data.