Self-Supervised Learning tasks have been able to produce very useful visual representations for downstream tasks like ImageNet classification. This video explains a study that attempts to combine self-supervised tasks to learn representations through learning multiple tasks at once! This result is very interesting!
Thanks for watching! Please Subscribe!
Paper Link: [ Ссылка ]
![](https://i.ytimg.com/vi/ODG60cYK7aU/maxresdefault.jpg)