Research map / Label-efficient and robust learning / Domain adaptation and generalization
Distribution Free Domain Generalization
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 |
|---|---|---|---|
| Eigenvalue (scaled MMD) standardization, DFDG-Eig with 1-NN, versus kernel DG baselines with 1-NN | kernel DG on synthetic data · Synthetic 6 cases | accuracy: DFDG-Eig 91.13 vs best baseline CIDG 91.03 (1-NN) | Helped |
| Using SVM classifier instead of 1-NN in DFDG (ablation) | DFDG-Eig · Synthetic 6 cases | accuracy: Case 1 92.97 vs 91.13; Case 6 71.23 vs 75.77 (1-NN better) | Mixed |
| Standardized DFDG (Eig or Cov) with 1-NN vs existing kernel DG methods on Office+Caltech and VLCS | DFDG 1-NN · Office+Caltech and VLCS | accuracy: similar performance to other DG methods | No clear effect |
| DFDG with SVM classifier vs other methods on Office+Caltech and VLCS | DFDG SVM · Office+Caltech and VLCS | accuracy: best in 9 of 16 cases | Helped |
| DFDG (1-NN) standardization applied to ERM ResNet-50 features on Terra Incognita | Terra Incognita · Terra Incognita (4 target domains) | accuracy: about 0.8% gain over ERM baseline (1-NN); 1.7% gain with SVM | Helped |
| Existing kernel DG methods (DICA, SCA, MDA, CIDG) applied to ERM features on Terra Incognita | kernel DG baselines · Terra Incognita | accuracy: failed to outperform ERM baseline | Hurt |
Closest papers by meaning
- Lost Domain Generalization Is a Natural Consequence of Lack of Training Domains (AAAI 2024 · similarity 0.89)
- Improving Domain Generalization with Domain Relations (ICLR 2024 · similarity 0.88)
- Robust Domain Generalization under Divergent Marginal and Conditional Distributions (NeurIPS 2026 · similarity 0.88)
- Set Valued Predictions For Robust Domain Generalization (ICML 2025 · similarity 0.87)
- Rethinking Multi-domain Generalization with A General Learning Objective (CVPR 2024 · similarity 0.87)
- Sharpness-Aware Model-Agnostic Long-Tailed Domain Generalization (AAAI 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.