AI Paper Atlas

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

Towards Novel Target Discovery Through Open-Set Domain Adaptation

Taotao Jing, Hongfu Liu, Zhengming Ding

ICCV 2021 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Source-free domain adaptation · 38 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 semantic attributes (full method vs visual-only variant) (ablation)Ours · D2AwA R->A open-set DAoverall accuracy OS: 89.2 -> 94.5Improved
add semantic attributes (full method vs visual-only variant) (ablation)Ours · I2AwA I->Aw open-set DAoverall accuracy OS: 48.7 -> 82.8Improved
add structure preserving partial alignment loss L^R (ablation)Ours · open-set DA seen categoriesaccuracyImproved
add binary seen/unseen classifier D (ablation)Ours · open-set DA and semantic recoveryaccuracyImproved
add attribute propagation over visual graph (ablation)Ours · semantic recovery unseen categoriesunseen accuracy UImproved
add visual-semantic fusion representation (ablation)Ours · open-set DA seen and semantic recovery unseenaccuracyImproved
use Ours instead of STA for unseen classesOurs vs STA · D2AwA P->A open-set DAunseen accuracy OS diamond: 5.7 -> 98.5Improved
use Ours instead of STA for unseen classesOurs vs STA · D2AwA R->A open-set DAunseen accuracy OS diamond: 7.2 -> 90.5Improved
Ours vs STA overall accuracyOurs vs STA · D2AwA P->A open-set DAoverall accuracy OS: 82.5 -> 81.0Hurt
Ours vs zero-shot methods on unseen classesOurs · D2AwA R->A semantic recoveryunseen accuracy U: 0.0 -> 37.8Improved
Ours without unseen attributes vs ABP* with unseen attributesOurs · D2AwA R->A semantic recoveryunseen accuracy U: 16.2 -> 37.8Improved

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