|Full Title:||Context-based Information Retrieval for E-science|
|Start Date:||28-Oct-2013 - 29-Oct-2013|
|Meeting Email:||click here to access email|
In the web of data, and especially e-science, heterogeneous resources such as data representation and description are potentially in relation. How can we formalize and exploit these relations? A possible answer is to consider contextual levels of representation (e.g., an ontology for research activity). Modeling these levels should bridge the gap between Information Science, Artificial Intelligence and Computer Science due to the use of standards like RDFS, OWL (and the Linked Data principles) and allow a convergence around the semantic definition of relations between resources. A primary focus of interest will concern the semantic characterization of these relations.
A foundational hypothesis is that e-science resources and publications must be considered in the context of their related resources. A corresponding e-science challenge could be the availability of resources corresponding to different steps of research activity. We are interesting both by the data context level or the specified relations between sets of descriptive (meta)data and the e-science level, or the user navigation between resources in a way to progress into its activity. In e-science, the context is defined using the relations between a published work and its used resources, including its community framework. These relations between heterogeneous data must be subsumed by a more general model of the research activity, involving knowledge structures like ontology, taxonomy and terminology.
Today, these constructions are tied to communities and are limited to a social categorization of the publication. How resources can deal with these communities to characterize frameworks and collective topics (and not only a more precise subject definition)? This question exceeds the usual profile characterization by the introduction of knowledge structures which also raises further issues such as: How the described content of a resource can be appropriate to describe some features of other related resources? What kind of information retrieval can be adapted to this new situation? How can we consider resource annotation and content extraction in this framework? What model of (scientific) activity is useful to govern the relations between these resources?
Relations between data structures can be specified in the ‘semantic web’ framework. But if we integrate the user, how the realized relations are stored and how they can be reused? How can we build communities of use founded on these practices? These questions introduce new perspectives about the record of the user behavior and then about web service elaboration. Furthermore, issues about resource content extraction and representation will reformulate the distinction between metadata based ontology and ontology for database structure.
The workshop will be a place for discussion between different perspectives on e-science development and a way to articulate its different parameters. It will foster the convergence between theoretical perspectives, social science analysis and more technical propositions.
|Linguistic Subfield:||Computational Linguistics|
|Calls and Conferences main page|