Learning from Unlabeled Videos for Recognition, Prediction, and Control

About this video

Hand-picked by a BrighterTube curator · found via round-3 back catalogue of a channel Lumen picks video by video · 82% confidence

Learning from unlabeled videos for recognition, prediction, and control is demonstrated. Long research talk covers self-supervised learning and representation learning from video.

A 83-minute video from Microsoft Research: “Learning from Unlabeled Videos for Recognition, Prediction, and Control”.

Video by Microsoft Research, played through YouTube's embedded player.

Part of Microsoft Vision+Language Summer Talk Series