AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

Big Self-Supervised Models Advance Medical Image Classification

Shekoofeh Azizi, Basil Mustafa, Fiona Ryan, Zachary Beaver, Jan Freyberg, Jonathan Deaton, Aaron Loh, Alan Karthikesalingam, Simon Kornblith, Ting Chen, Vivek Natarajan, Mohammad Norouzi

ICCV 2021 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Data-efficient and robust self-supervision · 723 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 in-domain unlabeled Derm pretraining on top of ImageNet pretraining (ablation)SimCLR ResNet-152 (2x) · Dermatologytop-1 accuracy (%): 66.38 -> 68.30Improved
add in-domain unlabeled CheXpert pretraining on top of ImageNet pretraining (ablation)SimCLR ResNet-152 (2x) · CheXpertmean AUC: 0.7671 -> 0.7729Improved
add MICLe multi-image positive pairs (ImageNet->Derm) (ablation)SimCLR ResNet-50 (4x) · Dermatologytop-1 accuracy (%): 67.63 -> 68.81Improved
add MICLe multi-image positive pairs (Derm only) (ablation)SimCLR ResNet-50 (4x) · Dermatologytop-1 accuracy (%): 66.93 -> 67.55Improved
add MICLe multi-image positive pairs (ImageNet->Derm) (ablation)SimCLR ResNet-152 (2x) · Dermatologytop-1 accuracy (%): 68.30 -> 68.43Improved
use partial instead of full augmentation in MICLe (Derm) (ablation)MICLe ResNet-152 (2x) · Dermatologytop-1 accuracy: 0.6697 -> 0.6761Improved
use partial instead of full augmentation in MICLe (ImageNet->Derm) (ablation)MICLe ResNet-152 (2x) · Dermatologytop-1 accuracy: 0.6928 -> 0.6889Hurt
replace supervised ImageNet pretraining with SimCLR ImageNet->DermResNet-152 (2x) · Dermatologytop-1 accuracy (%): 63.36 -> 69.43Improved
replace BiT ImageNet-21k supervised pretraining with MICLe ImageNet->DermResNet-152 (2x) vs ResNet-101 (3x) · Dermatologytop-1 accuracy (%): 68.45 -> 70.02Improved
replace BiT ImageNet-21k supervised pretraining with SimCLR ImageNet->CheXpertResNet-152 (2x) vs ResNet-101 (3x) · CheXpertmean AUC: 0.7720 -> 0.7729No clear effect
scale up the architecture from ResNet-50 (1x) to ResNet-152 (2x) with ImageNet->Derm pretrainingSimCLR · Dermatologytop-1 accuracy (%): 63.44 -> 68.30Improved
add in-domain CheXpert pretraining on top of ImageNet for ResNet-50 (1x) (ablation)SimCLR · CheXpertmean AUC: 0.7630 -> 0.7670No clear effect

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