Research map / Label-efficient and robust learning / Domain adaptation and generalization
Dynamic Transfer for Multi-Source Domain Adaptation
Check 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 |
|---|---|---|---|
| Replace static transfer with dynamic residual transfer (ablation) | MCD on DomainNet · DomainNet | accuracy: 38.5±0.61 -> 49.7±0.46 | Helped |
| Add self-training to DRT | DRT · DomainNet | accuracy: 49.7±0.46 -> 51.3±0.32 | Helped |
| Use channel attention instead of a static model (ablation) | Dynamic residual transfer · DomainNet | accuracy: 38.5±0.61 -> 47.8±0.59 | Helped |
| Use subspace routing instead of a static model (ablation) | Dynamic residual transfer · DomainNet | accuracy: 38.5±0.61 -> 49.7±0.46 | Helped |
| Combine channel attention and subspace routing (ablation) | Dynamic residual transfer · DomainNet | accuracy: 49.7±0.46 -> 49.6±0.50 | No clear effect |
| Use the combined dynamic residual for target infograph and quickdraw (ablation) | Subspace Routing DRT · DomainNet, target infograph and quickdraw | accuracy: 31.0±0.56 -> 31.6±0.61; 9.9±1.03 -> 11.9±0.96 | Helped |
| Use the combined dynamic residual on the other target domains (ablation) | Subspace Routing DRT · DomainNet, targets clipart, painting, real, and sketch | accuracy: 69.7±0.24 -> 69.1±0.35; 59.5±0.43 -> 58.2±0.25; 68.4±0.28 -> 67.8±0.36; 59.4±0.21 -> 58.8±0.44 | Hurt |
| Set the subspace routing basis count to K=4 rather than K=2, K=6, or K=8 (ablation) | Subspace Routing DRT · DomainNet | accuracy: 48.8, 49.7, 49.5, 49.3 | No clear effect |
| Add DRT without domain alignment loss (ablation) | Source Only · DomainNet | accuracy: 40.6±0.56 -> 42.8±0.74 | Helped |
| Add DRT while retaining ADDA alignment (ablation) | ADDA · DomainNet | accuracy: 32.2±0.63 -> 44.9±0.64 | Helped |
| Add DRT while retaining M³SDA-β alignment (ablation) | M³SDA-β · DomainNet | accuracy: 42.6±0.64 -> 48.8±0.55 | Helped |
| Use M³SDA-β alignment instead of MCD alignment with DRT (ablation) | DRT · DomainNet, target quickdraw | accuracy: 9.9±1.03 -> 13.6±0.34 | Helped |
Closest papers by meaning
- Revisiting the Domain Shift and Sample Uncertainty in Multi-source Active Domain Transfer (CVPR 2024 · similarity 0.91)
- Invertible Projection and Conditional Alignment for Multi-Source Blended-Target Domain Adaptation (AAAI 2025 · similarity 0.88)
- Cycle Self-Refinement for Multi-Source Domain Adaptation (AAAI 2024 · similarity 0.88)
- A General Representation-Based Approach to Multi-Source Domain Adaptation (ICML 2025 · similarity 0.88)
- Revisiting Source-Free Domain Adaptation: Insights into Representativeness, Generalization, and Variety (CVPR 2025 · similarity 0.87)
- Agile Multi-Source-Free Domain Adaptation (AAAI 2024 · 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.