HackerTube 中文?

Bay Area Vision Meeting: Unsupervised Feature Learning and Deep Learning

Google TechTalksYouTube ↗

Further reading 1

Where does learning get a signal without an answer key? →Start with an unlabeled patch of an image, then switch to a game that requires a sequence of actions. Neither has an answer key, but finding structure and choosing a move need different signals.This video helps answer · What can images teach without labels?

Caption notes

Learn representations from unlabeled sensory data before using scarce labels. Sparse coding on image patches illustrates the contrast with hand-designed features.

Original passages · 6
take all the unlabelled images, learn a better representation for images
Bay Area Vision Meeting: Unsupervised Feature Learning and Deep Learning · ▶ 11:12–11:22
The larger you can scale up these models, the better it does.
Bay Area Vision Meeting: Unsupervised Feature Learning and Deep Learning · ▶ 24:45–24:53
we don't understand the output of these algorithms actually is a problem.
Bay Area Vision Meeting: Unsupervised Feature Learning and Deep Learning · ▶ 39:47–39:58
we are essentially telling the algorithm which pixels are adjacent to which other input pixels.
Bay Area Vision Meeting: Unsupervised Feature Learning and Deep Learning · ▶ 47:12–47:21
find a better way to represent images than the raw input pixels
Bay Area Vision Meeting: Unsupervised Feature Learning and Deep Learning · ▶ 9:28–9:38
it learns a dictionary of bases functions
Bay Area Vision Meeting: Unsupervised Feature Learning and Deep Learning · ▶ 12:22–12:30

Hacker News comments

View original on HN ↗