AI Paper Atlas

Research map / Label-efficient and robust learning / Domain adaptation and generalization

Improve Unsupervised Pretraining for Few-Label Transfer

Suichan Li, Dongdong Chen, Yinpeng Chen, Lu Yuan, Lei Zhang, Qi Chu, Bin Liu, Nenghai Yu

ICCV 2021 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Conditional distribution alignment · 12 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add target-aware pretraining with unlabeled target dataMoCoV2 · 9 datasets mean, 1+9 labelsmean accuracy: 67.49 -> 74.15Improved
add progressive few-label transfer (ACKMeans eigen-sample annotation)TUP · 9 datasets mean, 1+9 labelsmean accuracy: 74.15 -> 76.69Improved
add progressive few-label transferTUP · 9 datasets mean, 1+1 labelsmean accuracy: 55.24 -> 59.30Improved
add sample re-balancing (20% target share) (ablation)TUP · DTD, 1+9 labelsaccuracy: 61.01 -> 66.58Improved
add sample re-balancing (20% target share) (ablation)TUP · DTD, 1+1 labelsaccuracy: 40.27 -> 44.86Improved
raise re-balancing ratio from 20% to 50% (ablation)TUP · DTD, 1+9 labelsaccuracy: 66.58 -> 63.86Hurt
continue unsupervised finetuning on target data instead of joint pretraining (ablation)MoCoV2 · DTD, 10 labels per classaccuracy: 58.68 -> 58.34No clear effect
unsupervised finetuning on target data instead of joint pretraining (ablation)MoCoV2 · STL10, 10 labels per classaccuracy: 90.85 -> 79.34Hurt
replace unsupervised finetuning with target-aware pretraining (ablation)MoCoV2 · Pet37, 1 label per classaccuracy: 40.01 -> 57.28Improved
add target data to pretraining, clustering qualityMoCoV2 · Food101cluster accuracy and transfer accuracy: 11.86 -> 39.23 cluster; 28.39 -> 52.25 transferImproved
use supervised ImageNet-1k pretraining vs ImageNet-100 (better clustering) (ablation)ResNet-50 pretrained · Pet, 5 labels per classtransfer accuracy: 45.65 -> 77.94Improved
use target-aware pretraining as initialization for FixMatchFixMatch · CIFAR10, 10 labels per classaccuracy: 92.15 -> 96.38Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy