AI Paper Atlas

Research map / Label-efficient and robust learning / Domain adaptation and generalization

Two at Once: Enhancing Learning and Generalization Capacities via IBN-Net

Xingang Pan, Ping Luo, Jianping Shi, Xiaoou Tang

ECCV 2018 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Augmentation and mixture-of-experts generalization · 967 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 half of BN with IN in shallow residual groups (IBN-a)ResNet50 · ImageNet valtop-1 error: 24.27 -> 22.54Improved
add IN after addition in early layers (IBN-b)ResNet50 · ImageNet val, Monet styletop-1 error: 54.51 -> 50.45Improved
add IBN-b to ResNet50 under std x1.5 transformResNet50 · ImageNet val, std x1.5top-1 error: 40.01 -> 23.64Improved
apply IBN-a to other backbonesDenseNet169, ResNet101, ResNeXt101, SE-ResNet101 · ImageNet valtop-1 error: 0.79%, 1.09%, 0.43%, 0.43%Improved
add IN to residual groups 1-3 vs none (ablation)IBN-Net50-a · ImageNet valtop-1 error: 24.27 -> 22.54Improved
add IN also to last residual group (group 4) (ablation)IBN-Net50-a · ImageNet valtop-1 error: 22.54 -> 22.96Hurt
raise IN channel ratio from 0.5 to 1 (ablation)IBN-Net50-a · ImageNet valtop-1 error: 22.54 -> 23.44Hurt
replace all BN with IN (ablation)ResNet50 · ImageNet valtop-1 error: 24.27 -> 28.56Hurt
add additional IBN layers (a x2) (ablation)IBN-Net50-a · ImageNet valtop-1 error: 22.54 -> 22.81No clear effect
combine IBN-a with IBN-d (ablation)IBN-Net50-d · ImageNet valtop-1 error: 22.86 -> 22.89No clear effect
IBN-a trained on Cityscapes, tested on GTA5ResNet50 · Cityscapes to GTA5mIoU: 29.4 -> 32.5Improved
IBN-b trained on Cityscapes, tested on GTA5ResNet50 · Cityscapes to GTA5mIoU: 29.4 -> 37.9Improved

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