AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Emerging Property of Masked Token for Effective Pre-training

Hyesong Choi, Hunsang Lee, Seyoung Joung, Hyejin Park, Jiyeong Kim, Dongbo Min

ECCV 2024 · Architectures and efficiency › Attention and transformers · approach: Token merging and masked pretraining · 10 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 masked token optimization to SimMIM and BootMAE pre-trainingSimMIM, BootMAE · ImageNet-1K ViT-B pre-training epochspre-training epochs to reach 800-epoch performance: 54% and 58% reductionImproved
add MTO to MAE and ConMIM pre-trainingMAE, ConMIM · ImageNet-1K ViT-Bpre-training epochs: 50% reductionImproved
add MTO to MAE with ViT-LMAE ViT-L · ImageNet-1Kpre-training epochs: 43% reductionImproved
use ranking loss alone without sparsity or entropy losses (ablation)SimMIM + MTO · ImageNet-1K 400 epochsTop-1 accuracy: 82.8 vs 83.2-83.3 for single other lossesHurt
add entropy loss L_e to ranking loss (ablation)SimMIM + MTO · ImageNet-1K 400 epochsTop-1 accuracy: 82.8 -> 83.4Improved
add all three losses instead of two (ablation)SimMIM + MTO · ImageNet-1K 400 epochsTop-1 accuracy: 83.4 -> 83.5Improved
add sparsity loss to SimMIM (ablation)SimMIM + MTO · ImageNet-1K 400 epochsTop-1 accuracy: 83.2 as sole lossImproved

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