|
@quocleix | |||||
|
Full comparison against state-of-the-art on ImageNet. Noisy Student is our method. Noisy Student + EfficientNet is 11% better than your favorite ResNet-50 😉 pic.twitter.com/BhwgJvSOYK
|
||||||
|
||||||
|
|
Quoc Le
@quocleix
|
12. stu |
|
Want to improve accuracy and robustness of your model? Use unlabeled data!
Our new work uses self-training on unlabeled data to achieve 87.4% top-1 on ImageNet, 1% better than SOTA. Huge gains are seen on harder benchmarks (ImageNet-A, C and P).
Link: arxiv.org/abs/1911.04252 pic.twitter.com/0umSnX7wui
|
||
|
||
|
|
Quoc Le
@quocleix
|
12. stu |
|
Example predictions on robustness benchmarks ImageNet-A, C and P. Black texts are correct predictions made by our model and red texts are incorrect predictions by our baseline model. pic.twitter.com/eem6tlfyPX
|
||
|
|
||
|
|
Quoc Le
@quocleix
|
12. stu |
|
Method is also super simple:
1) Train a classifier on ImageNet
2) Infer labels on a much larger unlabeled dataset
3) Train a larger classifier on the combined set
4) Iterate the process, adding noise
|
||
|
|
||
|
|
Quoc Le
@quocleix
|
13. stu |
|
I also highly recommend this nice video that explains the paper very well:
youtube.com/watch?v=Y8YaU9…
|
||
|
|
||
|
Jeremy Howard
@jeremyphoward
|
12. stu |
|
So grateful for all the really useful work you've been releasing, Quoc - I don't know how you do it! :)
I love this focus on making the most of data augmentation, pseudo-labeling, and other practically important techniques.
|
||
|
|
||
|
|
Quoc Le
@quocleix
|
13. stu |
|
Thanks @jeremyphoward!
|
||
|
|
||
|
Quantum Stat
@Quantum_Stat
|
12. stu |
|
@threadreaderapp unroll
|
||
|
|
||
|
Thread Reader App
@threadreaderapp
|
12. stu |
|
Hola please find the unroll here: Thread by @quocleix: "Want to improve accuracy and robustness of your model? Use unlabeled data! Our new work uses self-training on unlabeled […]" threadreaderapp.com/thread/1194334…
Share this if you think it's interesting. 🤖
|
||
|
|
||
|
Kuiyuan Yang
@kuiyuan
|
22. stu |
|
Worth to try this method on the whole index of Google Image Search. We did similar thing two years ago on whole Bing Image Search with some initial results reported in arxiv.org/abs/1708.08201.
|
||
|
|
||