hardmaru: Learning To Classify Images Without Labels
“On ImageNet, our approach can scale well up to 200 randomly selected classes, obtaining 69.3% top-1 and 85.5% top-5 accuracy, and marking a difference of less than 7.5% with fully-supervised methods.”
4 replies, 333 likes
roadrunner01: Learning To Classify Images Without Labels
abs: https://arxiv.org/abs/2005.12320 https://t.co/W4FIHB3dBJ
0 replies, 137 likes
Bindu Reddy 🔥❤️: Getting clean labeled data is a huge pain in ML
An exciting new work - "Learning to Classify Without Labels" applies self-supervised learning + clustering to the problem
Accuracy is comparable and only 7.5% < supervised methods!
1 replies, 31 likes
Wouter Van Gansbeke: Excited to share our work: “Learning To Classify Images Without Labels” w/ @svandenh1. Close to supervised performance and first to perform well on ImageNet for semantic clustering w/ no labels! Hope you enjoy it 😉.
1 replies, 19 likes
Carlos E. Perez: Use self-supervision to learn causal invariances and then apply clustering. https://arxiv.org/abs/2005.12320. .
1 replies, 6 likes
HotComputerScience: Most popular computer science paper of the day:
"Learning To Classify Images Without Labels"
0 replies, 3 likes
注目の最新arXiv【毎日更新】: 2020/05/25 投稿 2位
CV(Computer Vision and Pattern Recognition)
Learning To Classify Images Without Labels
9 Tweets 51 Retweets 178 Favorites
0 replies, 2 likes
Estesis: Learning To Classify Images Without Labels
Gansbeke et al.: https://arxiv.org/abs/2005.12320
#ArtificialIntelligence #DeepLearning #MachineLearning https://t.co/RXRkopbbNk
0 replies, 2 likes
Edward Dixon: Really extraordinary work here from @WGansbeke @svandenh1 @KU_Leuven @ETH_en & friends. Contrastive estimation makes an appearance. Especially interesting because they don't train end to end _and_ they avoid a dependency on low-level features. Stunning!
0 replies, 1 likes
Brundage Bot: Learning To Classify Images Without Labels. Wouter Van Gansbeke, Simon Vandenhende, Stamatios Georgoulis, Marc Proesmans, and Luc Van Gool http://arxiv.org/abs/2005.12320
1 replies, 0 likes
Found on Jun 02 2020 at https://arxiv.org/pdf/2005.12320.pdf