Research map / Label-efficient and robust learning / Domain adaptation and generalization
Exploiting Domain-Specific Features to Enhance 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 |
|---|---|---|---|
| add domain-specific features alongside domain-invariant ones (ablation) | DSDI · Background-Colored-MNIST | accuracy: 65.7 -> 89.7 | Helped |
| add domain-invariant features to domain-specific-only model (with meta-training) (ablation) | DSDI · Background-Colored-MNIST | accuracy: 75.3 -> 89.7 | Helped |
| add disentanglement (covariance) loss L_D (ablation) | DSDI · Background-Colored-MNIST | accuracy: 81.4 -> 89.7 | Helped |
| add meta-training on domain-specific representation (ablation) | DSDI · Background-Colored-MNIST | accuracy: 80.4 -> 89.7 | Helped |
| meta-train both Z_I and Z_S instead of only Z_S (ablation) | DSDI · Background-Colored-MNIST | accuracy: 89.7 -> 82.1 | Hurt |
| meta-train domain-invariant instead of domain-specific (ablation) | DSDI · Background-Colored-MNIST | accuracy: 89.7 -> 79.0 | Hurt |
| add meta-training to domain-specific-only model (ablation) | DS-only · Background-Colored-MNIST | accuracy: 70.7 -> 75.3 | Helped |
| add meta-training to domain-invariant-only model (ablation) | DI-only · Background-Colored-MNIST | accuracy: 65.7 -> 63.6 | Hurt |
| DSDI vs ERM across seven DG benchmarks | ResNet-50 / ConvNet · DomainBed average | accuracy: 66.6 -> 67.9 | Helped |
| DSDI vs best domain-invariant baseline CORAL | DomainBed · DomainBed average | accuracy: 67.5 -> 67.9 | Helped |
| DSDI vs ARM on Colored-MNIST (background-less) | DSDI · Colored-MNIST | accuracy: 56.2 -> 52.2 | Hurt |
Closest papers by meaning
- Improving Domain Generalization with Domain Relations (ICLR 2024 · similarity 0.91)
- Rethinking Multi-domain Generalization with A General Learning Objective (CVPR 2024 · similarity 0.90)
- Towards Generalizing to Unseen Domains with Few Labels (CVPR 2024 · similarity 0.90)
- Is Large-scale Pretraining the Secret to Good Domain Generalization? (ICLR 2025 · similarity 0.90)
- Unknown Domain Inconsistency Minimization for Domain Generalization (ICLR 2024 · similarity 0.89)
- Robust Domain Generalization under Divergent Marginal and Conditional Distributions (NeurIPS 2026 · 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.