Axe 1 - Cognitive Training and Cognitive Architectures - 31 janvier 2020

31-01-2020 - 10h00 à 12h00

The CHREST cognitive architecture

This presentation will provide an introduction to the CHREST cognitive architecture and will show how this architecture can account for a number of empirical data on the development of expertise and the acquisition of language. First, the components of the model will be presented, with a focus on its learning mechanisms, which are based on the notion of chunking. Then experiments on expertise in chess will be discussed, and it will be shown that computer simulations with CHREST can explain several key results. Simulations will also show that the architecture explains data on the development of language (e.g. acquisition of syntactic structures). Finally, the presentation will highlight the characteristics of CHREST allowing it to explain these empirical data. These characteristics include the idea of self-organisation, the emphasis on bounded rationality, the presence of a perception-learning-perception cycle, and the use of inputs representative of the domain to be learned.

Téléchargez l'affiche

Partager cet article :