AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Lightweight Vision Transformer with Bidirectional Interaction

Qihang Fan, Huaibo Huang, Xiaoqiang Zhou, Ran He

NeurIPS 2023 · Architectures and efficiency › Attention and transformers · approach: Efficient and robust vision transformers · 13 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
replace add+linear local-global fusion with bidirectional adaptive interaction (ablation)FAT-B0 · ImageNet-1KTop-1 accuracy (%): 76.2 -> 77.6Improved
replace cat+linear fusion with bidirectional interaction (ablation)FAT-B0 · ImageNet-1KTop-1 accuracy (%): 76.6 -> 77.6Improved
replace element-wise multiplication fusion with bidirectional interaction (ablation)FAT-B0 · ImageNet-1KTop-1 accuracy (%): 77.1 -> 77.6Improved
replace pooling downsampling with fine-grained downsampling (ablation)FAT-B0 · ImageNet-1KTop-1 accuracy (%): 77.2 -> 77.6Improved
replace overlapping large-stride conv with fine-grained downsampling (ablation)FAT-B0 · ImageNet-1KTop-1 accuracy (%): 77.3 -> 77.6Improved
add conditional positional encoding (ablation)FAT-B0 · ImageNet-1KTop-1 accuracy (%): 77.4 -> 77.6Improved
add local adaptive aggregation and bidirectional interaction to global-only (ablation)FAT-B0 · ImageNet-1KTop-1 accuracy (%): 76.2 -> 77.6Improved
add convolution stem (ablation)FAT-B0 · ImageNet-1KTop-1 accuracy (%): 77.2 -> 77.6Improved
add extra sigmoid activation in local aggregation (ablation)FAT-B0 · ImageNet-1KTop-1 accuracy (%): 77.6 -> 77.6No clear effect
increase DWConv kernel from 3 to staged 3,5,7,7 (ablation)FAT-B0 · ImageNet-1KTop-1 accuracy (%): 77.3 -> 77.6Improved
use uniform large kernel 9 in all stages vs staged kernels (ablation)FAT-B0 · ImageNet-1KTop-1 accuracy (%): 77.6 -> 77.4Hurt
replace Max-SA with FASA on same backboneFAT-B0 · ImageNet-1KTop-1 accuracy (%): 75.7 -> 77.6Improved

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