Sesei: A CG-Based Filter for Internet Search Engines

The web faces a major contradiction: it has been created to present information to human beings but the increase of available information sources on the web makes it more difficult for people to find the most relevant documents when browsing through numerous pages of URLs returned by search engines. Classic indexing and retrieval techniques have an important limitation: the retrieved documents do not necessarily match the meaning of the user’s query. In order to address this issue, we designed a system capable of matching conceptual structures extracted from both the user’s query and documents retrieved by a classic search engine like Google. This article presents our fully functional application and provides details about its behavior. The application provides a new framework for broader experiments in the manipulation of conceptual structures extracted from natural language documents and is built upon research and tools developped by the CG community.