AI Paper Atlas

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

Contrastive Learning with Stronger Augmentations

Xiao Wang, Guo-Jun Qi

TPAMI 2022 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Inductive bias and minimal representations · 293 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 stronger augmentation with DDM loss to MoCo v2MoCo v2 · ImageNet linear eval (200 epochs)top-1 accuracy: 67.5 -> 69.4Improved
add multiple stronger augmentations (multi-crop)CLSA · ImageNet linear eval (200 epochs)top-1 accuracy: 69.4 -> 73.3Improved
use strong query instead of weak query in MoCo v2 (ablation)MoCo v2 · ImageNet linear eval (200 epochs)top-1 accuracy: 67.5 -> 67.7No clear effect
use strong query and strong key in MoCo v2 (ablation)MoCo v2 · ImageNet linear eval (200 epochs)top-1 accuracy: 67.5 -> 67.0Hurt
add DDM loss instead of plain contrastive loss (ablation)CLSA · ImageNet linear eval (200 epochs)top-1 accuracy: 68.0 -> 69.4Improved
train CLSA longer (100 to 200 epochs) (ablation)CLSA · ImageNet linear evaltop-1 accuracy: 67.2 -> 69.4Improved
use CLSA at equal running time versus MoCo v2CLSA · ImageNet linear evaltop-1 accuracy: 67.5 -> 68.3Improved
add stronger augmentation strength s=5 versus none (ablation)CLSA · ImageNet KNN (100 epochs, K=20)KNN accuracy: 51.9 -> 53.0Improved
increase strength s from 5 to 7 (ablation)CLSA · ImageNet KNN (100 epochs, K=20)KNN accuracy: 53.0 -> 52.6No clear effect
add CLSA to BYOLBYOL · ImageNet linear eval (200 epochs)top-1 accuracy: 66.5 -> 67.3Improved
add CLSA to MoCo v2 (generalization)MoCo v2 · ImageNet linear eval (200 epochs)top-1 accuracy: 67.5 -> 69.4Improved
train longer with multi-crop CLSA (800 epochs)CLSA* · ImageNet linear evaltop-1 accuracy: 76.2 vs 76.5Improved

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