AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Delving Deep Into the Generalization of Vision Transformers Under Distribution Shifts

Chongzhi Zhang, Mingyuan Zhang, Shanghang Zhang, Daisheng Jin, Qiang Zhou, Zhongang Cai, Haiyu Zhao, Xianglong Liu, Ziwei Liu

CVPR 2022 · Architectures and efficiency › Attention and transformers · approach: Efficient and robust vision transformers · 104 citations (OpenAlex)

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 prototype-based self-supervision and mutual information maximizationDeiT-B/16 · DomainNetAvg. accuracy: 46.07 -> 51.07Improved
Add domain-adversarial trainingDeiT-S/16 · DomainNetAvg. accuracy: 43.22 -> 46.13Improved
Add minimax entropy trainingDeiT-S/16 · DomainNetAvg. accuracy: 43.22 -> 47.00Improved
Remove mutual information maximization from T-SSL (ablation)DeiT-S/16 T-SSL · DomainNetAvg. accuracy: 47.37 -> 46.35Hurt
Remove in-domain self-supervision from T-SSL (ablation)DeiT-S/16 T-SSL · DomainNetAvg. accuracy: 47.37 -> 47.30No clear effect
Use the complete T-SSL combination rather than either loss alone (ablation)DeiT-S/16 · DomainNetAvg. accuracy: 47.30, 46.35 -> 47.37Improved

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