Machine Learning Street Talk (MLST) cover image

#55 Self-Supervised Vision Models (Dr. Ishan Misra - FAIR).

Machine Learning Street Talk (MLST)

00:00

Advancements in Self-Supervised Learning

This chapter explores the evolution of self-supervised learning architectures in computer vision, transitioning from traditional models to innovative approaches like ResNet and contrastive learning. It discusses the significance of techniques such as Polyak averaging and the challenges of mode collapse, while emphasizing the importance of fine-tuning similarity tasks and optimizing model performance. Additionally, the conversation touches on the interplay between different modalities and the implications for language translation, showcasing the intersection of ideas across fields.

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