Research map / Label-efficient and robust learning / Domain adaptation and generalization
Progressive Feature Alignment for Unsupervised 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 |
|---|---|---|---|
| Use progressive easy-to-hard selection instead of random target selection (ablation) | PFAN · A→W; I→P; SVHN→MNIST | accuracy (%): 77.0 -> 81.9; 67.0 -> 68.0; 87.2 -> 92.5 | Helped |
| Select progressively rather than use all target samples throughout training (ablation) | PFAN · A→W; I→P; SVHN→MNIST | accuracy (%): 77.0 -> 81.9; 67.0 -> 68.0; 87.2 -> 92.5 | Mixed |
| Add adaptive prototype alignment (ablation) | PFAN · A→W; I→P; SVHN→MNIST | accuracy (%): 76.4 -> 83.0; 67.1 -> 68.5; 82.0 -> 93.9 | Helped |
| Use global accumulated prototypes rather than current-mini-batch prototypes (ablation) | PFAN · A→W; I→P; SVHN→MNIST | accuracy (%): 82.2 -> 83.0; 68.1 -> 68.5; 93.0 -> 93.9 | Mixed |
| Add temperature scaling to the source classifier (ablation) | PFAN · A→W; I→P; SVHN→MNIST | accuracy (%): 80.6 -> 83.0; 67.9 -> 68.5; 92.1 -> 93.9 | Helped |
| Use full PFAN configuration rather than source-only training (ablation) | PFAN · A→W; I→P; SVHN→MNIST | accuracy (%): 61.6 -> 81.9; 66.2 -> 68.0; 60.1 -> 92.5 | Helped |
| Use category-level alignment in addition to domain alignment | PFAN vs RevGrad · Office-31 and ImageCLEF-DA | accuracy (%) | Helped |
| Apply PFAN rather than the strongest compared baseline | PFAN · Office-31; ImageCLEF-DA; digit classification | accuracy (%): 79.1% -> 80.4% | Mixed |
| Use PFAN rather than MSTN on MNIST→SVHN | PFAN · MNIST→SVHN | accuracy (%): +4.8% | Helped |
Closest papers by meaning
- Make the U in UDA Matter: Invariant Consistency Learning for Unsupervised Domain Adaptation (NeurIPS 2023 · similarity 0.91)
- Co-regularized Alignment for Unsupervised Domain Adaptation (NeurIPS 2018 · similarity 0.91)
- Structure-Aware Feature Fusion for Unsupervised Domain Adaptation (AAAI 2020 · similarity 0.91)
- Exploiting Local Feature Patterns for Unsupervised Domain Adaptation (AAAI 2019 · similarity 0.90)
- Can We Evaluate Domain Adaptation Models Without Target-Domain Labels? (ICLR 2024 · similarity 0.90)
- CASUAL: Conditional Support Alignment for Domain Adaptation with Label Shift (AAAI 2025 · similarity 0.90)
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.