AI Paper Atlas

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

Learning Unbiased Transferability for Domain Adaptation by Uncertainty Modeling

Jian Hu, Haowen Zhong, Fei Yang, Shaogang Gong, Guile Wu, Junchi Yan

ECCV 2022 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Conditional distribution alignment · 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
add UTEPDANN · Office-31 (UDA)avg accuracy: 86.1 -> 88.0Improved
add UTEPMDD · Office-31 (UDA)avg accuracy: 88.9 -> 89.9Improved
add UTEPCADA · Office-31 (UDA)avg accuracy: 89.5 -> 90.0Improved
add UTEPTransPar · Office-31 (UDA)avg accuracy: 89.9 -> 90.8Improved
add UTEPDANN · Office-Home (UDA)avg accuracy: 65.2 -> 67.3Improved
add UTEPCADA · Office-Home (UDA)avg accuracy: 70.2 -> 70.6Improved
add UTEPDANN · VisDA-2017mean accuracy: 72.3 -> 79.2Improved
add UTEPMDD · VisDA-2017mean accuracy: 81.4 -> 82.3Improved
add UTEP rather than UPS pseudo labelingDANN · Office-Home (3-shot SSL)avg accuracy: 63.3 -> 65.1Improved
add bias loss to the full UTEP (ablation)DANN+UTEP · Office-31 (UDA)avg accuracy: 87.2 -> 88.0Improved
add negative cross-entropy to the full UTEP (ablation)DANN+UTEP · Office-31 (UDA)avg accuracy: 87.3 -> 88.0Improved
add target bias loss (TBL) to the full UTEP (ablation)DANN+UTEP · Office-31 (UDA)avg accuracy: 87.3 -> 88.0Improved

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