Research map / Label-efficient and robust learning / Domain adaptation and generalization
Transferable Semantic Augmentation for 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 |
|---|---|---|---|
| Add TSA to BSP | BSP · Office-Home UDA | average accuracy: 66.3 -> 71.2 | Helped |
| Add TSA to BSP | BSP · Office-31 UDA | average accuracy: 88.5 -> 90.6 | Helped |
| Add TSA to BSP | BSP · VisDA-2017 Synthetic→Real | accuracy: 76.9 -> 82.0 | Helped |
| Add TSA to JADA | JADA · Digital Datasets UDA | average accuracy: 97.4 -> 98.5 | Helped |
| Add inter-domain mean difference to the TSA sampling distribution (ablation) | DANN+TSA · Office-31 UDA | average accuracy: 86.9 -> 89.4 | Helped |
| Add target intra-class covariance to the TSA sampling distribution (ablation) | DANN+TSA · Office-31 UDA | average accuracy: 87.5 -> 89.4 | Helped |
| Add mutual information loss to TSA (ablation) | DANN+TSA · Office-31 UDA | average accuracy: 87.2 -> 89.4 | Helped |
| Use mutual information loss alone without TSA semantic augmentation (ablation) | DANN · Office-31 UDA | average accuracy: 82.2 -> 86.4 | Hurt |
| Replace TSA with ISDA (ablation) | DANN · Office-31 UDA | average accuracy: 89.4 -> 85.4 | Hurt |
| Use a memory module instead of iterative estimation for class statistics (ablation) | TSA · Office-31 UDA | accuracy: 86.5 -> 89.3 | Helped |
| Increase target training data from 80% to 100% (ablation) | ResNet-50+TSA · Office-31 UDA | average accuracy: 89.1 -> 89.3 | No clear effect |
| Add TSA to CDAN | CDAN · Digital Datasets UDA | accuracy on MNIST→USPS: 95.6±0.1 -> 95.1±0.2 | Hurt |
Closest papers by meaning
- Transferable Adversarial Training: A General Approach to Adapting Deep Classifiers (ICML 2019 · similarity 0.89)
- Balancing Discriminability and Transferability for Source-Free Domain Adaptation (ICML 2022 · similarity 0.88)
- A General Representation-Based Approach to Multi-Source Domain Adaptation (ICML 2025 · similarity 0.88)
- SF(DA)$^2$: Source-free Domain Adaptation Through the Lens of Data Augmentation (ICLR 2024 · similarity 0.87)
- SimProF: A Simple Probabilistic Framework for Unsupervised Domain Adaptation (AAAI 2025 · similarity 0.87)
- Batch Singular Value Polarization and Weighted Semantic Augmentation for Universal Domain Adaptation (ICML 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.