Contrastive Learning Insights
Yann discusses the concept of collapse in neural networks and the evolution of contrastive learning techniques. He highlights the significance of negative samples and the challenges posed by high dimensional spaces. The conversation also touches on innovative approaches like BYOL, which eliminates the need for negative samples, showcasing the rapid advancements in the field.In this clip
From this podcast
The Gradient
Yann LeCun on his Start in Research and Self-Supervised Learning
Related Questions