AI Paper Atlas

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

Hierarchical Unsupervised Relation Distillation for Source Free Domain Adaptation

Bowei Xing, Xianghua Ying, Ruibin Wang, Ruohao Guo, Ji Shi, Wenzhen Yue

ECCV 2024 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Source-free domain adaptation · 4 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 local clustering relation distillation (ablation)SHOT baseline · VisDA-2017average per-class accuracy: 82.9 -> 87.8Improved
add global semantic relation distillation (ablation)SHOT baseline · VisDA-2017average per-class accuracy: 82.9 -> 87.0Improved
combine local and global relations (ablation)HRD · VisDA-2017average per-class accuracy: 87.8 -> 89.0Improved
add adaptive ensemble re-weighting (ablation)HRD · Office-Homeaverage accuracy: 74.7 -> 75.2Improved
add adaptive ensemble re-weighting (ablation)HRD · DomainNetaverage accuracy: 68.9 -> 69.9Improved
add source-like/target-specific data division (ablation)HRD · VisDA-2017average per-class accuracy: 87.3 -> 89.8Improved
add class-balanced division threshold (ablation)HRD · DomainNetaverage accuracy: 69.4 -> 69.9Improved
use relation distillation instead of logit distillation (ablation)HRD · VisDA-2017accuracy: 85.2 -> 89.8Improved
use teacher-student logit distillation instead of hard pseudo labels (ablation)SFDA self-training · VisDA-2017accuracy: 82.6 -> 85.2Improved
increase neighbors K beyond 5 (ablation)HRD · VisDA-2017average accuracyHurt
vary EMA coefficient among 0.95, 0.99, 0.999 (ablation)HRD · VisDA-2017average accuracy: above 89.0%No clear effect
add HRD to SHOT baselineSHOT · Office-Homeaverage accuracy: 71.8 -> 75.2Improved

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