Machine Learning Street Talk (MLST) cover image

ICLR 2020: Yann LeCun and Energy-Based Models

Machine Learning Street Talk (MLST)

00:00

Understanding Energy-Based Models

This chapter explores energy-based models and their integration with factor graphs, highlighting the importance of multimodal predictions. It discusses the role of energy functions in optimizing outputs and their implications for discrete versus continuous data, particularly in the context of sampling challenges. The chapter also contrasts conditional and unconditional energy-based models, using examples like K-means clustering to illustrate key concepts.

Transcript
Play full episode

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app