Thomas: There’s a misconception that you need Google sized datasets and millions of dollars of computing power to do deep learning. You can use transfer learning to take an existing model and use on a smaller dataset. #lfosls
Replying to @linuxfoundation
And/or use cheaper, higher-level *weak* supervision- see recent progress (http://lld-workshop.github.io ) and work w/ @HazyResearch on using domain expertise as weak supervision (http://snorkel.stanford.edu ) as examples. Good complement to transfer learning!
0 replies
1 retweet
7 likes
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.