Research map / Label-efficient and robust learning / Domain adaptation and generalization
Progressive Graph Learning for Open-Set Domain Adaptation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 changed | Where | Measured | Effect |
|---|---|---|---|
| apply Progressive Graph Learning instead of the strongest reported baseline | Office-Home, averaged across 12 transfer pairs · Office-Home | OS*: 70.4 -> 76.1 | Helped |
| apply Progressive Graph Learning instead of the strongest reported baseline | Syn2Real-O · Syn2Real-O | OS*: 52.2 -> 66.8 | Helped |
| apply Progressive Graph Learning instead of the strongest reported baseline | VisDA-17 · VisDA-17 | OS*: 63.9 -> 82.8 | Helped |
| add progressive learning rather than training in one step (ablation) | PGL on Syn2Real-O · Syn2Real-O | OS: 54.4 -> 65.5 | Helped |
| use focal loss rather than NLL loss for node classification (ablation) | PGL on Syn2Real-O · Syn2Real-O | OS: 56.9 -> 65.5 | Helped |
| add GNNs for class-wise manifold learning (ablation) | PGL on Syn2Real-O · Syn2Real-O | OS: 57.8 -> 65.5 | Helped |
| add mix-up of pseudo-labeled target samples into training episodes (ablation) | PGL on Syn2Real-O · Syn2Real-O | OS: 62.5 -> 65.5 | Helped |
| reduce the pseudo-label enlarging factor from 0.20 to 0.05 (ablation) | PGL on Syn2Real-O · Syn2Real-O | OS*: 63.3 -> 66.5 | Helped |
| reduce the pseudo-label enlarging factor from 0.20 to 0.05 (ablation) | PGL on Office-Home (Ar→Cl) · Office-Home (Ar→Cl) | OS*: 61.1 -> 63.1 | Helped |
| remove edge supervision by setting its coefficient μ to zero (ablation) | PGL on Office-Home (Ar→Cl) · Office-Home (Ar→Cl) | OS: drops by 4.6% | Hurt |
| remove adversarial learning by setting its coefficient γ to zero (ablation) | PGL on Office-Home (Ar→Cl) · Office-Home (Ar→Cl) | OS: drops by 10.2% | Hurt |
| increase openness on Syn2Real-O across the tested levels (ablation) | PGL · Syn2Real-O | OS | No clear effect |
Closest papers by meaning
- LEARNING FACTORIZED REPRESENTATIONS FOR OPEN-SET DOMAIN ADAPTATION (ICLR 2019 · similarity 0.89)
- AdaGraph: Unifying Predictive and Continuous Domain Adaptation Through Graphs (CVPR 2019 · similarity 0.88)
- Towards Unsupervised Open-Set Graph Domain Adaptation via Dual Reprogramming (NeurIPS 2025 · similarity 0.88)
- Open-Set Graph Domain Adaptation via Separate Domain Alignment (AAAI 2024 · similarity 0.87)
- GCAN: Graph Convolutional Adversarial Network for Unsupervised Domain Adaptation (CVPR 2019 · similarity 0.87)
- Towards Inheritable Models for Open-Set Domain Adaptation (CVPR 2020 · similarity 0.87)
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.