Université de Montréal

Institute for Advanced Study School of Mathematics seminar investigating symplectic rigidity, Floer homology persistence modules, and invariant subsets of Hamiltonian flows.

Yoshua Bengio outlines a research trajectory from intuitive, data-driven System 1 deep learning towards architectures capable of conscious reasoning and causal understanding.

Reviews distributed representations and attention mechanisms before proposing sparse, modular causal factor graphs for out-of-distribution generalization.

A panel discussion featuring leading researchers on the intersection of AI, cognitive science, and the philosophical implications of large language models.