AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Vicinity Vision Transformer

Weixuan Sun, Zhen Qin, Hui Deng, Jianyuan Wang, Yi Zhang, Kaihao Zhang, Nick M. Barnes, Stan T. Birchfield, Lingpeng Kong, Yiran Zhong

TPAMI 2023 · Architectures and efficiency › Attention and transformers · approach: Efficient and robust vision transformers · 66 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 2D vicinity locality (ablation)VVT-Tiny · CIFAR-100top-1 accuracy (%): 82.13 -> 82.92Improved
add 2D vicinity locality (ablation)VVT-Tiny · ImageNet-1ktop-1 accuracy (%): 78.0 -> 79.4Improved
use 1D locality (cosFormer) instead of no locality (ablation)VVT-Tiny · CIFAR-100top-1 accuracy (%): 82.13 -> 76.78Hurt
use 1D locality (cosFormer) in 2D vision data (ablation)VVT-Tiny · ImageNet-1ktop-1 accuracy (%): failed to convergeHurt
raise feature reduction ratio from 2 to 8 (ablation)VVT · CIFAR-100top-1 accuracy (%): 82.92 -> 80.60Hurt
use feature reduction ratio 1 instead of 2 (ablation)VVT · CIFAR-100top-1 accuracy (%): 82.89 vs 82.92No clear effect
raise feature reduction ratio from 2 to 4 (ablation)VVT · CIFAR-100top-1 accuracy (%): 82.92 -> 82.00Hurt
add feature preserving connection (FPC) (ablation)VVT-Tiny · ImageNet-1ktop-1 accuracy (%): 77.9 -> 79.4Improved
add FRA and FPC to Performer attention (ablation)VVT-T with Performer · ImageNet-1ktop-1 accuracy (%): 77.2 -> 77.3No clear effect
use vanilla softmax attention instead of vicinity attention (ablation)VVT-T · ImageNet-1ktop-1 accuracy (%): 79.4 -> 80.0Improved
use vicinity attention instead of vanilla attention (ablation)VVT-T · CIFAR-100top-1 accuracy (%): 82.68 -> 82.92Improved
use vicinity attention instead of Performer (ablation)VVT-T · ImageNet-1ktop-1 accuracy (%): 77.2 -> 79.4Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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