In this classic lecture, Co-Founder and Chief Scientist Shane Legg explores how researchers chart improvements in AI and what model of intelligence we should develop.
Research Scientist Nando de Freitas speaks about training large neural networks and the process of building tools that learn to solve many tasks with few data each.
Petar Veličković tells how graph neural networks (GNNs) were derived from first principles and explains how they emerged from several lines of research.