Représentation de connaissances complexes : Un formalisme à base de rôles sémantiques

Abstract : The work presented in this thesis deals with knowledge representation. It is integrated in the project AMICAL (Interactive Multi-agents Architecture Fellow in Learning to Read) whose purpose is the development of an aided learning to read environment. This study focus on the search of a semantic representation of texts used in learning sessions. The problematics is to find a knowledge representation formalism which is capable of answering to a particular functional situation while taking into account various kinds of constraints. Three possibilities were presented : retain an existing knowledge representation formalism, make existing formalisms cooperate or propose a new formalism. The third solution has been taken after showing difficulties met by existing formalisms in filling some formulated constraints. We propose in this thesis a knowledge representation system based on knowledge structures named DCAS (Composite Description With Meaning). They are the basic units of reasoning. These structures are based on a set of various sorts of cognitive entities. We describe the different roles played by these cognitive entities inside a DCAS, as well as the representationnal and inferential capacities of a system based on these notions.
Document type :
Complete list of metadatas

Cited literature [22 references]  Display  Hide  Download
Contributor : Paul Lotin <>
Submitted on : Monday, November 15, 2010 - 11:31:29 AM
Last modification on : Friday, March 29, 2019 - 12:10:03 AM
Long-term archiving on : Wednesday, February 16, 2011 - 2:47:51 AM


  • HAL Id : tel-00536060, version 1



Paul Lotin. Représentation de connaissances complexes : Un formalisme à base de rôles sémantiques. Interface homme-machine [cs.HC]. Université Blaise Pascal - Clermont-Ferrand II, 1995. Français. ⟨tel-00536060⟩



Record views


Files downloads