Objective-Driven AI: Towards AI Systems That Can Learn, Remember, Reason, and Plan

Join a lecture by renowned AI researcher Yann LeCun from New York University and META as he explores the future of artificial intelligence systems capable of human-like learning and reasoning. Delve into the limitations of current AI architectures like Auto-Regressive Large Language Models and discover a proposed modular cognitive architecture that could revolutionize machine learning. Learn about the Hierarchical Joint Embedding Predictive Architecture (H-JEPA), a predictive world model that enables systems to anticipate action consequences and optimize objectives while maintaining safety and controllability. Explore how machines could potentially acquire common sense, learn efficiently like humans and animals, and develop reasoning and planning capabilities through self-supervised learning techniques that create abstract representations of percepts designed to be both maximally informative and predictable.