Section: Application Domains

Semantic networks and knowledge representation

Actors from the industry are now forced to store and filter huge amounts of information in digital formats on subjects concerning their domain of activity. This state of affairs calls for the development of technologies capable of helping human users to explore, filter and organize information and moreover discover concepts from non-structured or semi-structured data. After we had recognized here a situation where graph visualization could help, we grouped with industrial actors in a technology transfer effort. Although the industry can certainly benefit from our actual expertise, the project nevertheless requires genuine research.

Our research in this field is conducted in close collaboration with the industry and focuses on developing techniques and tools devoted to competitive and strategic watch. After a successful experience working with PIKKO and AMI Software(See the URLs http://www.pikko-software.com and http://www.amisw.com .) through a 2 year ANR RNTL project, we recently grouped our efforts with Thalès Communications and Xerox. The TANGUY(TANGUY stands for “From Text to Arguments through Networks with Goals and User Initiative”.) project is funded through the ANR CONTINT call. This project is concerned with processing legal text documents, mixed with unstructured and uncertain information gathered from various sources into a semantic network. The project focuses on developing a genuine approach to help legal experts to explore dense legal documents.

Our efforts focus on building and delivering answers from visual queries on a semantic network. We also plan to provide tools to visually support reporting activities. The project requires the design of astute graphical representations and data handling.

We see in this project a real opportunity to address issues raised within the knowledge representation domain. In a sense, this project should reinforce our expertise on the processing and visualization of dynamic data.