FIT-IT Lasso - Lookup & Alignment Service with Semantic Open data

Project LASSO intends to deploy, improve and extend Linked Data (LD) infrastructure in three different use cases. Central feature of all use cases is a lookup service which helps to augment already existing, formalized knowledge with facts from the Linked Open Data (LOD) cloud. These three use cases are Semantic Desktop, Enterprise Collaboration and Inspiration Services. Currently available systems in all of these application domains are sparsely using data from the Semantic Web. All systems rely on their local knowledge repositories and make no use from publicly available data like from DBpedia or other LD sources. Augmenting local knowledge repositories with additional facts from the web can improve several knowledge services in all usage scenarios in a significant way three examples: (1) Finding collaborators within an enterprise platform who will provide the right information in a given context will be more efficient. (2) Search for similar documents or resources can be done with more background information, thus will improve significantly. (3) Generating inspiring ideas from a given node can be supported by more concise associations. The crucial point in all knowledge systems built around the linked data cloud is the ability to identify resources linkable with local resources. Current lookup services are mainly based on string matching (using labels of resources), although promising approaches from other disciplines (like ontology alignment, graph matching, spreading activation, etc.) might improve such a service. The proposed project will significantly enhance the quality of mappings between local concepts and LOD concepts thus enabling the faster uptake of LOD for the benefit of the individual user as well as a group of users by developing two novel services: (i) a Context Elicitation Service including automatic context detection, context model population as well as contextual query formulation mechanisms and (ii) a LD Lookup Service including sophisticated algorithms for mapping local concepts to LD concepts, automatic query alignment and query execution as well as novel search and rank algorithms. The three proofs of concept implementations in the use case partners domains will prove the feasibility of the LASSO approach. Since LD sources evolve dynamically and quite often offer inconsistent information, detecting valuable information from the LOD cloud is like catching a fast moving horse with a LASSO .