C-MANTC

The projet called « C-Mantic » aims at providing tools for the search of scientific information on the web, using the methods of text-based semantics in addition to the classical search methods (based on keywords). Hence, the project does aim at realizing new search engines but at completing search engines with more precise tools, with respect to the meaning, the type, the genre of returned documents. Thanks to project the it will be also possible to formalise robust and efficient methodologies using textbased semantics and to select adequate software tools. The projects has two levels: − At the level of search engines: which methods and tools are adequate for searching scientific corpora (human sciences and exact sciences) ? − At the level of the design of the tools, and of the processing of the corpora: which methodologies and which tools are relevant for extracting the linguistic and semantic clues that are necessary for the semantic-enabled search engines ? The project also has three dimensions: − Methodology (research) − Technology (corpus processing tools) − Applications (search engines for scientific documents) ... and 3 languages: French, English and Chinese (at experimental level) The topic addressed will be the fight against the use of tobacco (medecine, pharmacology and public opinion). Besides the multitude of Web documents relevant to the topic, there are several advantages in choosing this field: in particular, the scope of discourses (medical, journalistic, ads...), the large reusability of the approach for similar domains of public interest, the economic dimension.