Project: Spatial Ontology Language for multimedia Information Modeling

The goal of the project is to extend the ontology web language OWL to support effective storage and reasoning on spatial information, and to demonstrate the power of such extension for automatic processing of textual and graphical information._x000D__x000D_The technological problem this project is proposed to solve is that the current technological tools for description of semantic knowledge are insufficient to support adequate automated reasoning on the inherent spatial properties of concepts. Properties as dimension, orientation and location can only be dealt with in an ad-hoc manner whilst these are among the basic properties of physical concepts expressed in many ontologies. Instead of adding these properties every time they are required their definition should be supported by the ontology language and reasoning on these properties should be supported from a very low level. This would enable any application using the new ontology language to become “spatially aware” and perform spatial reasoning and consistency checking as part of the basic information modelling process. This project sets out to develop such a new ontology language together with the required reasoning tools, and will demonstrate the power of this approach in two proof of concept applications. To demonstrate the fundamental impact of this project, the applicators are selected from the opposites of the information landscape: free text translation and image search. _x000D_The translation POC is aimed at improving the semantic understanding by implementing spatial consistency checking. The measure for success is an improved translation quality._x000D_The image search POC is aimed at improving automatic image classification and tagging by implementing spatial scale and orientation checking. Th measure for success is an improved search result. _x000D__x000D_This project is to expand the current state of the art ontology languages (OWL, RDF-S, RDF) and their support for automated reasoning by adding a spatial dimension to improve real time context aware information analysis. This technology will provide ontology based (semantic) information systems to venture beyond a static world and add the concepts of space and change. _x000D__x000D_Current ontology languages have limited expressive power for describing spatial knowledge. Information with a spatial component can be described using an ontology by treating locations as ordinary concepts. However, by doing so the temporal-spatial consequences about the described events (locations and movement) is lost in the formalisation. This effectively means that knowledge about the spatial aspects and allowable positioning and changes in space (orientation, scale, location and travel of concepts) cannot be efficiently described inside the ontology, although it is valid and persistent knowledge about the doCO._x000D__x000D_Reasoning on changes occurring in space is limited to statics without the notion of time for describing real world situations. Dynamic spatial restrictions are defined in view of the time in which they occur. Therefore this project will ultimately connect to TOWL (IST-2004-26896) In TOWL temporal knowledge expression is introduced into the standard web ontology language._x000D__x000D_The CO result of this project is improved semantic modelling of real word data. This improves the quality of the semantic model (the “understanding”) of extracted data by an automated system. The prevention of spatial semantic inconsistencies will directly improve the quality of any system that is currently using ontology based knowledge. Immediate applications are natural language or image based information systems (search engines, matching- or classification systems, news alert systems, image analysis systems etc.)_x000D__x000D_The project will be a joint effort by a well balanced consortium of European SME's. SemLab of The Netherlands as an recognised innovator of semantic web technology and CO technology P of the mentioned TOWL project will be the CO driving force for the base technology development. Primarily the definition of the web language and the development of the reasoner will be their CO task. AGM-Labs of Turkey will be supporting this development together with the creation of the actual ontologies for use in the development of the proof of concept applications. AGM lab has since its founding in 2004 established a firm base in knowledge management in general and semantic information mapping in particular. The reCOing two Ps Tilde of Lithuania and PicSearch of Sweden are CO European companies in their particular field of natural language tools and image classification and retrieval respectively. For this reason they are perfectly geared to take up the technological advances and implement them in their semantic framework. In the project they will closely work with the previous Ps, but are COly selected for their ability to implement and assess the actual market value of the technological developments in the limited time frame the project allows.

Acronym SOLIM (Reference Number: 4365)
Duration 01/09/2008 - 01/03/2011
Project Topic This project is to improve context aware information analysis by expanding the state of the art ontology languages and their support for automated reasoning by adding a spatial dimension. This will provide semantic systems to venture beyond a static world and add the concepts of space and change.
Project Results
(after finalisation)
The CO result of the SOLIM project is the extension of the ontology web language (OWL) to support the effective storage of and reasoning on the spatial information. The achieved result has demonstrated the potential of such an extension for the automatic processing of textual and graphical information in two Proof of Concept (POC) applications: Translation POC and Image Search POC correspondingly._x000D__x000D_Tilde IT was leading the translation work package (WP4) and was in charge of the Translation POC design, development and implementation. It was an innovative research resulting in the English-Lithuanian statistical factored phrase-based machine _x000D_translation (MT) system with a spatial ontology. The system is based on the Moses toolkit and is enriched with semantic knowledge inferred from the spatial ontology. The ontology was developed on the basis of the GeoNames (http://www.geonames.org/) database (more than 15 000 toponyms), implemented in the web ontology language (OWL), and integrated into the MT process. Spatial knowledge was added as an additional factor in the statistical translation model and used for toponym disambiguation during MT. The implemented MT approach was evaluated against the baseline system without spatial knowledge. A multifaceted evaluation strategy including automatic metrics, human evaluation and linguistic analysis, was implemented to perform evaluation experiments. The results of the evaluation have shown the improvement in the output quality of MT with spatial knowledge (Skadinš et al. 2011). _x000D__x000D_Skadinš R., Gornostay T., Šics V. Toponym Disambiguation in an English-Lithuanian SMT System with Spatial Knowledge. In the Proceedings of the 18th Nordic Conference on Computational Linguistics (NODALIDA 2011), May 11-13, 2011, University of Latvia, Riga, Latvia._x000D_
Network Eurostars
Call Eurostars Cut-Off 1

Project partner

Number Name Role Country
4 AGMLab Information Technologies Partner Turkey
4 Tilde informacines technologijos Partner Lithuania
4 Picsearch AB Partner Sweden
4 ZooRobotics BV Coordinator Netherlands