AI Paper Atlas

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

Self-Supervised Visual Representations Learning by Contrastive Mask Prediction

Yucheng Zhao, Guangting Wang, Chong Luo, Wenjun Zeng, Zheng-Jun Zha

ICCV 2021 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Masked autoencoder pretraining · 43 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 mask prediction head (3 blocks vs none) (ablation)MaskCo · ImageNet linear conv5 (pretrained on ImageNet)top-1 accuracy: 41.9 -> 63.1Improved
add mask prediction head (3 blocks vs none) (ablation)MaskCo · Pascal VOC detection (pretrained on ImageNet)AP: 56.1 -> 56.0No clear effect
increase MPH from 1 to 3 blocks (ablation)MaskCo · ImageNet linear conv5 (pretrained on ImageNet)top-1 accuracy: 59.0 -> 63.1Improved
increase MPH from 3 to 4 blocks (ablation)MaskCo · ImageNet linear conv5 (pretrained on ImageNet)top-1 accuracy: 63.1 -> 63.0No clear effect
add masking of the query region (ablation)MaskCo without MPH · Pascal VOC detection (pretrained on ImageNet)AP: 53.6 -> 56.1Improved
add masking of the query region (ablation)MaskCo without MPH · ImageNet linear conv4 (pretrained on ImageNet)top-1 accuracy: 51.9 -> 57.2Improved
add masking of the query region (ablation)MaskCo without MPH · ImageNet linear conv5 (pretrained on ImageNet)top-1 accuracy: 52.3 -> 41.9Hurt
add masking of the query region (ablation)MaskCo without MPH · Pascal VOC detection (pretrained on COCO)AP: 54.1 -> 56.2Improved
add intra-image negatives to inter-image negatives (ablation)MaskCo · Pascal VOC detection (pretrained on ImageNet)AP: 56.0 -> 56.4Improved
add intra-image negatives to inter-image negatives (ablation)MaskCo · ImageNet linear conv5 (pretrained on ImageNet)top-1 accuracy: 63.1 -> 61.3Hurt
use only intra-image negatives instead of inter-image (ablation)MaskCo · ImageNet linear conv5 (pretrained on ImageNet)top-1 accuracy: 63.1 -> 48.6Hurt
replace MoCo v2 with MaskCo as pretraining method on COCOCOCO pretraining · ImageNet linear conv5 (pretrained on COCO)top-1 accuracy: 51.4 -> 56.7Improved

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