AI Paper Atlas

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

Rethinking the Role of Pre-Trained Networks in Source-Free Domain Adaptation

Wenyu Zhang, Li Shen, Chuan-Sheng Foo

ICCV 2023 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Source-free domain adaptation · 27 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 co-learning with Swin-B pre-trained branchsource-trained ResNet-50 · Office-Homeavg accuracy: 59.6 -> 83.5Improved
add co-learning with ResNet-50 ImageNet branch to SHOT++SHOT++ · Office-31avg accuracy: 89.2 -> 89.1No clear effect
add co-learning with Swin-B to NRCNRC · Office-Homeavg accuracy: 72.0 -> 81.4Improved
add co-learning with Swin-B to SHOTSHOT · VisDA-Cavg accuracy: 82.4 -> 85.2Improved
use MatchOrConf instead of SelfConf pseudolabeling (ablation)co-learning · Office-Home (A->C,A->P,A->R)avg accuracy: 67.0 -> 77.7Improved
use MatchOrConf instead of MatchAndConf (ablation)co-learning · Office-Home (A->C,A->P,A->R)avg accuracy: 75.4 -> 77.7Improved
lower confidence threshold gamma from 0.9 to 0.1 (ablation)co-learning · VisDA-Caccuracy: 85.5 -> 87.1Improved
raise confidence threshold gamma from 0.1 to 0.7 (ablation)co-learning · Office-Homeaccuracy: 76.1 -> 77.5Improved
finetune only the classifier in pre-trained branch vs finetuning feature extractor (ablation)co-learning · Office-Home (A->C,A->P,A->R)avg accuracy: 76.2 -> 77.7Improved
estimate class centroids from all samples rather than pseudolabeled only (ablation)co-learning · Office-Home (A->C,A->P,A->R)avg accuracy: 77.4 -> 77.7Improved
iteratively refine class centroids (3 iterations) (ablation)co-learning · Office-Home (A->C,A->P,A->R)avg accuracy: 77.7 -> 76.7Hurt
adapt ResNet-50 with Swin-B co-learning vs fixed Swin-B with source classification head (ablation)Swin-B · Office-Home (A->C,A->P,A->R)avg accuracy: 81.0 -> 83.5Improved

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