AI Paper Atlas

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

Uncertainty-Guided Source-Free Domain Adaptation

Subhankar Roy, Martin Trapp, Andrea Pilzer, Juho Kannala, Nicu Sebe, Elisa Ricci, Arno Solin

ECCV 2022 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Source-free domain adaptation · 83 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 Laplace-approximation uncertainty-weighted entropy loss to SHOT-IMSHOT-IM · Office-Home closed-setavg accuracy (%): 70.5 -> 71.8Improved
weight entropy loss with MAP-model certainty instead of no weighting (ablation)SHOT-IM · Office-Home closed-setavg accuracy (%): 70.5 -> 71.2Improved
replace MAP weights with Laplace-approximation (posterior) weights (ablation)SHOT-IM + entropy weighting · Office-Home closed-setavg accuracy (%): 71.2 -> 71.8Improved
add uncertainty-guided weighting to SHOT-IMSHOT-IM · Office-Home open-setOS accuracy (%): 71.5 -> 73.5Improved
add uncertainty-guided weighting to SHOT-IMSHOT-IM · VisDA-Caccuracy (%): 80.3 -> 81.2Improved
add uncertainty-guided weighting to SHOT-IMSHOT-IM · DomainNet (avg)accuracy (%): 21.98 -> 26.13Improved
add uncertainty-guided weighting to SHOT-IM, source quickdrawSHOT-IM · DomainNet (source quickdraw)accuracy (%): 10.76 -> 10.44Hurt
add uncertainty-guided weighting to SHOT-IMSHOT-IM · CIFAR9 to STL9accuracy (%): 75.69 -> 77.04Improved
add uncertainty-guided weighting to SHOT-IMSHOT-IM · Office31 (saturated)avg accuracy (%): 87.3 -> 88.6Improved
add nearest-centroid pseudo-labelling of SHOT to U-SFANU-SFAN · Office-Home closed-setavg accuracy (%): 71.8 -> 71.9No clear effect
add nearest-centroid pseudo-labelling of SHOT to U-SFANU-SFAN · VisDA-Caccuracy (%): 81.2 -> 82.7Improved
add nearest-centroid pseudo-labelling of SHOT to U-SFANU-SFAN · Office31avg accuracy (%): 88.6 -> 88.8No 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