Research map / Label-efficient and robust learning / Domain adaptation and generalization
Domain Generalization by Learning and Removing Domain-specific Features
Yu Ding, Lei Wang, Bin Liang, Shuming Liang, Yang Wang, Fang Chen
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 |
|---|---|---|---|
| Explicitly remove learned domain-specific image features using LRDG | PACS (AlexNet) · PACS Avg. | accuracy: 76.05 -> 77.33 | Improved |
| Explicitly remove learned domain-specific image features using LRDG | PACS (ResNet18) · PACS Avg. | accuracy: 85.15 -> 85.48 | Improved |
| Explicitly remove learned domain-specific image features using LRDG | PACS (ResNet50) · PACS Avg. | accuracy: 88.10 -> 88.63 | Improved |
| Explicitly remove learned domain-specific image features using LRDG | VLCS (AlexNet) · VLCS Avg. | accuracy: 75.43 -> 75.15 | Hurt |
| Explicitly remove learned domain-specific image features using LRDG | Office-Home (ResNet18) · Office-Home Avg. | accuracy: 65.50 -> 65.75 | Improved |
| Use entropy uncertainty loss instead of least-likely loss (ablation) | LRDG domain-specific classifiers · Source-domain classification randomness | classification randomness | Improved |
| Map source and target images to remove domain-specific features | PACS domain pairs · Pairwise source-domain divergence | Proxy A-distance | Improved |
| Map source and target images to remove domain-specific features | PACS source-target pairs · Closest mixture source-target domain pairs | Proxy A-distance | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SWAD — SWAD: Domain Generalization by Seeking Flat Minima (NeurIPS 2021) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Episodic Training for Domain Generalization (ICCV 2019 · similarity 0.91)
- Deeper, Broader and Artier Domain Generalization (ICCV 2017 · similarity 0.90)
- Deep Domain Generalization via Conditional Invariant Adversarial Networks (ECCV 2018 · similarity 0.90)
- Learning to Balance Specificity and Invariance for In and Out of Domain Generalization (ECCV 2020 · similarity 0.89)
- Unlearning during Training: Domain-Specific Gradient Ascent for Domain Generalization (ICLR 2026 · similarity 0.89)
- Domain Generalization With Adversarial Feature Learning (CVPR 2018 · 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.
