AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Deep Reinforced Attention Learning for Quality-Aware Visual Recognition

Duo Li, Qifeng Chen

ECCV 2020 · Architectures and efficiency › Attention and transformers · approach: Sparse and structured attention · 9 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 DREAL recurrent critic supervision of attentionSE-ResNet-50 · ImageNet valtop-1 error (%): 22.616 -> 22.152Improved
add DREAL recurrent critic supervision of attentionCBAM-ResNet-50 · ImageNet valtop-1 error (%): 22.386 -> 21.802Improved
add DREAL recurrent critic supervision of attentionSRM-ResNet-50 · ImageNet valtop-1 error (%): 22.700 -> 22.348Improved
add DREAL recurrent critic supervision of attentionSE-ResNet-101 · ImageNet valtop-1 error (%): 21.488 -> 20.732Improved
add DREAL recurrent critic supervision of attentionCBAM-ResNet-101 · ImageNet valtop-1 error (%): 21.518 -> 20.682Improved
add DREAL recurrent critic supervision of attentionSRM-ResNet-101 · ImageNet valtop-1 error (%): 21.404 -> 20.474Improved
add DREAL criticCBAM-ResNet-50 · ImageNet valtop-5 error (%): 6.172 -> 6.084Improved
add DREAL criticResNet-50 with spatial and channel attention · Market-1501Rank-1 / mAP (%): 94.7 -> 96.1 Rank-1, 85.9 -> 89.6 mAPImproved
add DREAL criticResNet-50 with spatial and channel attention · DukeMTMC-reIDRank-1 / mAP (%): 86.4 -> 89.6 Rank-1, 76.2 -> 79.8 mAPImproved
add attention modules to the ReID baselineResNet-50 · Market-1501Rank-1 (%): 93.5 -> 94.7Improved
use deeper ResNet-101 instead of ResNet-50 with DREALSE/CBAM/SRM · ImageNet valtop-1 error gain: 22.152 -> 20.732 for SEImproved
use more complicated recurrent critic architectures (ablation)DREAL critic · ImageNet valtop-1 errorNo clear effect

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