AI Paper Atlas

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

Prior Knowledge Guided Unsupervised Domain Adaptation

Tao Sun, Cheng Lu, Haibin Ling

ECCV 2022 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Feature alignment for unsupervised adaptation · 35 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 unary-bound prior knowledge (sigma=0) rectificationSHOT · Office-Home RS-UTaccuracy: 63.2 -> 66.6Improved
add unary-bound prior knowledge (sigma=0) rectificationSHOT · DomainNetaccuracy: 79.1 -> 81.0Improved
add unary-bound prior knowledge (sigma=0)SHOT · VisDA-2017accuracy: 82.9 -> 86.1Improved
add binary-relationship prior knowledgeSHOT · Office-Home RS-UTaccuracy: 63.2 -> 66.2Improved
add unary-bound prior knowledge (sigma=0)SHOT · Office-Home PDAaccuracy: 79.3 -> 86.8Improved
loosen bound width sigma from 0.0 to 2.0 (ablation)kSHOT · Office-Home RS-UTaccuracy: 66.6 -> 63.3Hurt
add unary-bound prior knowledge (sigma=0)DINE · Office-Homeaccuracy: 69.7 -> 71.4Improved
replace adaptive label smoothing with vanilla smoothingDINE · Office-Homeaccuracy: 69.7 -> 68.4Hurt
use smoothed labels instead of one-hot labels (ablation)kDINE · Office-Homeaccuracy: 70.7 -> 71.2Improved
add smooth regularization (ablation)kSHOT UB sigma=0.0 · Office-Home RS-UTaccuracy: 66.3 -> 66.6Improved
estimate class prior from 0.5% of target data instead of full data (ablation)kSHOT · VisDA-2017accuracy: 86.13 -> 84.06Hurt
combine UB sigma=0.5 with BR (ablation)kSHOT · VisDA-2017accuracy: 83.9 -> 84.4Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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