AI Paper Atlas

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

Exploring Robustness of Unsupervised Domain Adaptation in Semantic Segmentation

Jinyu Yang, Chunyuan Li, Weizhi An, Hehuan Ma, Yuzhi Guo, Yu Rong, Peilin Zhao, Junzhou Huang

ICCV 2021 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Adaptation for semantic segmentation · 36 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
train with adversarial self-supervision (ASSUDA) on top of BDLBDL · GTA5 to Cityscapes VGG16, eps=0.25mIoU: 14.6 -> 32.8Improved
add ASSUDA to BDLBDL · GTA5 to Cityscapes VGG16, eps=0.5mIoU: 7.1 -> 19.1Improved
add ASSUDA to BDLBDL · GTA5 to Cityscapes ResNet101, eps=0.25mIoU: 19.9 -> 33.1Improved
add ASSUDA to BDLBDL · SYNTHIA to Cityscapes VGG16, eps=0.25mIoU: 13.6 -> 33.6Improved
add ASSUDA to BDLBDL · SYNTHIA to Cityscapes ResNet101, eps=0.5mIoU: 11.0 -> 11.7No clear effect
attack BDL with FGSM eps=0.1 from surrogate PSPNetBDL VGG16 · GTA5 to CityscapesmIoU: 41.3 -> 30.5Hurt
increase FGSM eps to 0.5 on BDLBDL ResNet101 · GTA5 to CityscapesmIoU: 48.5 -> 6.50Hurt
use perturbation magnitude eps_m=1 instead of 0.1 (ablation)ASSUDA · GTA5 to Cityscapes VGG16, eps=0.25mIoU: 21.2 -> 32.8Improved
use eps_m=1 instead of 0.1 (ablation)ASSUDA · GTA5 to Cityscapes VGG16, eps=0.5mIoU: 8.3 -> 19.1Improved
use eps_m=1 instead of 0.1 (ablation)ASSUDA · SYNTHIA to Cityscapes ResNet101, eps=0.1mIoU: 45.5 -> 48.5Improved
use eps_m=1 instead of 0.1 (ablation)ASSUDA · SYNTHIA to Cityscapes VGG16, eps=0.5mIoU: 8.9 -> 18.5Improved
train on adversarial examples with VGG16 on mild perturbationASSUDA · SYNTHIA to Cityscapes VGG16, eps=0.1mIoU drop: -0.8Improved

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