AI Paper Atlas

Research map / Label-efficient and robust learning / Semi-supervised learning

Semi-Supervised Domain Generalization with Known and Unknown Classes

Lei Zhang, Ji-Fu Li, Wei Wang

NeurIPS 2023 · Label-efficient and robust learning › Semi-supervised learning · approach: Safe and realistic semi-supervised learning · 2 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 unsupervised loss on known classes with fixed thresholds to supervised loss (ablation)CWAEE · OfficeHomeknown-class accuracy: 68.13 -> 72.30Improved
add unsupervised loss on known classes with fixed thresholds (ablation)CWAEE · OfficeHomeunknown-class AUROC: 69.73 -> 64.47Hurt
add unsupervised loss on unknown classes with fixed thresholds (ablation)CWAEE · OfficeHomeunknown-class AUROC: 64.47 -> 65.16Improved
add unsupervised loss on unknown classes with fixed thresholds (ablation)CWAEE · OfficeHomeknown-class accuracy: 72.30 -> 72.38No clear effect
replace fixed thresholds with class-wise adaptive thresholds (ablation)CWAEE · OfficeHomeunknown-class AUROC: 65.16 -> 74.44Improved
replace fixed thresholds with class-wise adaptive thresholds (ablation)CWAEE · OfficeHomeknown-class accuracy: 72.38 -> 73.70Improved
add consistency regularization loss (ablation)CWAEE · OfficeHomeknown-class accuracy: 73.70 -> 74.34Improved
add consistency regularization loss (ablation)CWAEE · OfficeHomeunknown-class AUROC: 74.44 -> 75.20Improved
use CWAEE instead of OpenMatchCWAEE · PACSknown-class accuracy: 77.04 -> 84.20Improved
use CWAEE instead of StyleMatchCWAEE · miniDomainNetunknown-class AUROC: 62.97 -> 74.35Improved
use CWAEE instead of FixMatch for seen unknown AUROCCWAEE · PACSseen unknown-class AUROC: 53.84 -> 84.09Improved
increase labels per class from 5 to 20CWAEE · OfficeHomeknown-class accuracy: 71.24 -> 76.09Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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