University of Toronto

Geoffrey Hinton·University of Toronto

Contrasts immortal weight-sharing digital models with mortal analog biological computation—and examines knowledge distillation across heterogeneous architectures.

Daniel Litt·University of Toronto

An exploration of non-abelian Hodge theory, developing structures on fundamental group representations analogous to those in cohomology, and presenting new related conjectures.

David Duvenaud reflects on Neural Ordinary Differential Equations, a foundational contribution presented at NeurIPS 2019, discussing its implications and future directions.

Daniel Litt·University of Toronto

Addressing the question of when algebraic differential equations yield algebraic function solutions, tracing the history from Fuchs to contemporary conjectures by Grothendieck and Katz.

Costas Sarris·University of Toronto

Costas Sarris presents recent advancements in applying scientific machine learning techniques to complex electromagnetic field computations, exploring time-domain modeling and wave propagation.

Patterson Hume&Donald Ivey·University of Toronto

A foundational exploration of inertial and non-inertial frames of reference, presented by University of Toronto faculty in this classic 1960 educational film.