Research map / Label-efficient and robust learning / Domain adaptation and generalization
Model Adaptation: Unsupervised Domain Adaptation Without Source Data
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 |
|---|---|---|---|
| Add generated target-style labeled samples to the adaptation training (ablation) | 3C-GAN model adaptation · SVHN→MNIST | accuracy (%): 68.1±1.5 -> 97.9±0.2 | Helped |
| Add weight regularization that keeps adapted weights near source-model weights (ablation) | 3C-GAN model adaptation · MNIST→MNIST-M | accuracy (%): 91.8±0.5 -> 94.2±0.3 | Helped |
| Add clustering-based regularization to generated-sample loss and weight regularization (ablation) | 3C-GAN model adaptation · MNIST→MNIST-M | accuracy (%): 94.2±0.3 -> 97.0±0.1 | Helped |
| Include local smoothness regularization (ablation) | 3C-GAN model adaptation · A→W | accuracy (%): 93.4±0.3 -> 93.7±0.2 | Helped |
| Adapt using 3C-GAN instead of the Source-Only model | Source-Only model · MNIST→MNIST-M | accuracy (%): 54.2±0.9 -> 98.5±0.2 | Helped |
| Adapt using 3C-GAN instead of the Source-Only model | Source-Only model · Office-31 average | accuracy (%): 76.1 -> 89.6 | Helped |
| Use the enhanced ResNet101 backbone | Our Model on VisDA17 · VisDA17 validation | class mean accuracy (%): 81.6 -> 83.3 | Helped |
| Remove the generated-sample loss (ablation) | 3C-GAN model adaptation · Digit adaptation tasks | accuracy (%) | Hurt |
Closest papers by meaning
- Universal Source-Free Domain Adaptation (CVPR 2020 · similarity 0.90)
- Co-regularized Alignment for Unsupervised Domain Adaptation (NeurIPS 2018 · similarity 0.90)
- Unsupervised Pixel-Level Domain Adaptation With Generative Adversarial Networks (CVPR 2017 · similarity 0.90)
- Unsupervised Multi-Source Domain Adaptation Without Access to Source Data (CVPR 2021 · similarity 0.89)
- Can We Evaluate Domain Adaptation Models Without Target-Domain Labels? (ICLR 2024 · similarity 0.89)
- A Prototype-Oriented Framework for Unsupervised Domain Adaptation (NeurIPS 2021 · similarity 0.89)
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.