Research map / Label-efficient and robust learning / Domain adaptation and generalization
Self-Paced Collaborative and Adversarial Network for Unsupervised Domain Adaptation
Weichen Zhang, Dong Xu, Wanli Ouyang, Wen Li
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add image-classifier-based pseudo-label sample selection (CSS) to CAN (ablation) | CAN · Office-31 (ResNet50) | accuracy: 82.4 -> 88.1 | Improved |
| add domain-classifier-based sample selection (DSS) on top of CSS (ablation) | SPCAN w/o DSS · Office-31 (ResNet50) | accuracy: 88.1 -> 89.1 | Improved |
| add CSS when only DSS is used (ablation) | SPCAN w/o CSS · Office-31 (ResNet50) | accuracy: 82.4 -> 89.1 | Improved |
| apply self-paced sample selection to DANN (ablation) | DANN · Office-31 (ResNet50) | accuracy: 80.9 -> 86.2 | Improved |
| learn block weights lambda_l instead of fixing them to 1/3 (ablation) | SPCAN · Office-31 (ResNet50) | accuracy: 87.1 -> 89.1 | Improved |
| use positive (domain-specific) fixed weights at shallow blocks versus negative (ablation) | sSPCAN · Office-31 (ResNet50) | accuracy: 85.9 -> 87.1 | Improved |
| add collaborative learning with multiple domain classifiers to DANN | DANN · Office-31 (ResNet50) | accuracy: 80.9 -> 82.4 | Improved |
| swap backbone from DANN to SPCAN with VGG19 | DANN · Office-31 (VGG19) | accuracy: 78.4 -> 87.5 | Improved |
| swap DANN to SPCAN with DenseNet161 backbone | DANN · Office-31 (DenseNet161) | accuracy: 82.7 -> 90.1 | Improved |
| add collaborative and adversarial scheme to DANN for video | DANN · UCF101-10/HMDB51-10 | accuracy: 87.7 -> 90.4 | Improved |
| add self-paced sample selection to CAN for video | CAN · UCF101-10/HMDB51-10 | accuracy: 90.4 -> 92.1 | Improved |
| exchange pseudo-labeled samples between RGB and flow streams | SPCAN · UCF101-10/HMDB51-10 | accuracy: 92.1 -> 93.5 | Improved |
Closest papers by meaning
- Collaborative and Adversarial Network for Unsupervised Domain Adaptation (CVPR 2018 · similarity 0.93)
- Unsupervised Video Domain Adaptation with Masked Pre-Training and Collaborative Self-Training (CVPR 2024 · similarity 0.89)
- Beyond Sharing Weights for Deep Domain Adaptation (TPAMI 2018 · similarity 0.88)
- Unsupervised Video Domain Adaptation for Action Recognition: A Disentanglement Perspective (NeurIPS 2023 · similarity 0.88)
- Domain-Symmetric Networks for Adversarial Domain Adaptation (CVPR 2019 · similarity 0.88)
- Domain-Adversarial Training of Neural Networks (JMLR 2016 · similarity 0.88)
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.
