AI Paper Atlas

Research map / Probabilistic and scientific modeling / Causal discovery

Causal Attention for Unbiased Visual Recognition

Tan Wang, Chang Zhou, Qianru Sun, Hanwang Zhang

ICCV 2021 · Probabilistic and scientific modeling › Causal discovery · approach: Causal representation identifiability · 136 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 the Maxi-Game partition update instead of rebooting training each phase (ablation)CNN-CaaM · NICO testaccuracy: 44.46 -> 46.62Improved
add learned partition update instead of randomizing theta each phase (ablation)CNN-CaaM · NICO testaccuracy: 43.38 -> 46.62Improved
add learned partition update instead of randomizing theta each phase (ablation)ViT-CaaM · NICO testaccuracy: 36.15 -> 37.61Improved
add progressive weight retention instead of rebooting training (ablation)ViT-CaaM · NICO testaccuracy: 35.46 -> 37.61Improved
use 2 CaaM blocks instead of 1 (ablation)CNN-CaaM · NICO testaccuracy: 44.54 -> 45.77Improved
use 4 CaaM blocks instead of 2 (ablation)CNN-CaaM · NICO testaccuracy: 45.77 -> 45.31No clear effect
use m=4 splits instead of m=2 (ablation)CNN-CaaM · NICO testaccuracy: 44.92 -> 45.77No clear effect
use m=8 splits instead of m=4 (ablation)CNN-CaaM · NICO testaccuracy: 45.77 -> 45.74No clear effect
remove human-annotated partitions for CaaMCNN-CaaM · ImageNet-Aaccuracy: 35.60 -> 38.55Improved
remove human-annotated partitions for IRMIRM (CNN) · ImageNet-Aaccuracy: 33.52 -> 33.39No clear effect
add CaaM versus conventional attentionResNet18+CBAM · NICO testaccuracy: 42.46 -> 46.62Improved
add CaaM versus conventional attention (attention accuracy)ResNet18+CBAM · ImageNet-9 attention accuracyattention accuracy: 69.73 -> 78.37Improved

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