Research map / Label-efficient and robust learning / Domain adaptation and generalization
Boosting Domain Adaptation by Discovering Latent Domains
Read the paperCheck 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 |
|---|---|---|---|
| Discover latent source domains and align their features separately | Domain adaptation on MNIST-MNIST-m to USPS · MNIST-MNIST-m to USPS | accuracy: 81.7 -> 82.5 | Helped |
| Use known source-domain labels instead of discovering them | Domain adaptation on MNIST-MNIST-m to USPS · MNIST-MNIST-m to USPS | accuracy: 82.5 -> 84.2 | Helped |
| Vary the discovered source-domain count to k=3 (ablation) | Ours on MNIST-MNIST-m to USPS · MNIST-MNIST-m to USPS | accuracy: 82.5 -> 82.2 | No clear effect |
| Vary the discovered source-domain count to k=4 (ablation) | Ours on MNIST-MNIST-m to USPS · MNIST-MNIST-m to USPS | accuracy: 82.5 -> 82.7 | No clear effect |
| Vary the discovered source-domain count to k=5 (ablation) | Ours on MNIST-MNIST-m to USPS · MNIST-MNIST-m to USPS | accuracy: 82.5 -> 82.4 | No clear effect |
| Discover latent source domains rather than unify all sources | PACS domain adaptation · PACS mean | accuracy: 84.2 -> 85.3 | Helped |
| Discover latent source domains rather than unify all sources | Office-31 domain adaptation · Office-31 mean | accuracy: 83.5 -> 83.6 | No clear effect |
| Use known source-domain labels instead of discovering them | Office-31 domain adaptation · Office-31 mean | accuracy: 83.6 -> 84.5 | Helped |
| Discover latent source domains | Office-31 AlexNet-feature comparison · Office-31, W-D to A | accuracy: 48.9 -> 64.2 | Helped |
| Discover latent source domains | Office-Caltech domain adaptation · Office-Caltech mean | accuracy: 89.8 -> 91.8 | Helped |
Closest papers by meaning
- Visual Representation Learning over Latent Domains (ICLR 2022 · similarity 0.93)
- Domain Conditioned Adaptation Network (AAAI 2020 · similarity 0.90)
- Domain Generalization Using a Mixture of Multiple Latent Domains (AAAI 2020 · similarity 0.90)
- A General Representation-Based Approach to Multi-Source Domain Adaptation (ICML 2025 · similarity 0.90)
- Domain Adaptation with Invariant Representation Learning: What Transformations to Learn? (NeurIPS 2021 · similarity 0.89)
- Multi-Source Distilling Domain Adaptation (AAAI 2020 · 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.