Le vendredi 1er novembre 2024, de 10h00 à 12h00, Mariam Thalos (University of Tennessee) donnera une conférence intitulée « Reasoning in Context ».

L’événement aura lieu dans la salle 738 du bâtiment Leacock, situé au 855 rue Sherbrooke, à l’Université McGill. Pour y assister via Zoom, c’est ici.

La conférence est organisée par le Groupe de recherche interuniversitaire sur la normativité (GRIN).

Résumé

The concept of intelligence has been difficult to get one’s arms around. Surprisingly, the same can be said also of the notion of reasoning. This talk aims at shedding some light on certain aspects of human reasoning – reasoning for practical life. This will put us in a better position to make some comparisons between (some aspects of) human reasoning, and what Large Language Models (LLMs) are doing – they look to be doing quite different things.  Reasoning, at least as humans do it, involves architecture that Computer Science has apparently abandoned for the current generation of AI models. By articulating an account of the machinery required for reasoning, we will be able to pose questions about how human reasoning operates collective settings as well as in individual settings.