ICLR 2020: Yann LeCun and Energy-Based Models
Machine Learning Street Talk (MLST)
00:00
Understanding Energy-Based Models and Latent Variables
This chapter explores the complexities of energy-based models, emphasizing the role of latent variables in simplifying inference processes. It discusses practical applications through clustering models and draws parallels between energy functions and decision-making in strategic games. The conversation further delves into high-dimensional manifolds, training methods, and the nuances of leveraging energy functions in machine learning.
Transcript
Play full episode
