Research map / Label-efficient and robust learning / Domain adaptation and generalization
Bridging Domain Invariance and Diversity: A Fine-Grained Risk Bound for Domain Generalization
Xi Wang, Liang Bai, Xian Yang, Richard Yi Da Xu, Jiye Liang
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 |
|---|---|---|---|
| add MixStyle on all three source domains (generate 3 new domains) (ablation) | ResNet-18 ERM · Digits-DG (leave-one-domain-out) | average accuracy: 62.29% -> 78.11% | Improved |
| add MixStyle on all three source domains (generate 3 new domains) (ablation) | ResNet-18 ERM · PACS (leave-one-domain-out) | average accuracy: 75.14% -> 79.33% | Improved |
| add MixStyle on all three source domains (generate 3 new domains) (ablation) | ResNet-18 ERM · Office-Home (leave-one-domain-out) | average accuracy: 61.91% -> 62.76% | Improved |
| add MixStyle on two source domains (1 new domain) (ablation) | ResNet-18 ERM · Digits-DG (leave-one-domain-out) | average accuracy: 62.29% -> 71.91% | Improved |
| add MixStyle on two source domains (1 new domain) (ablation) | ResNet-18 ERM · PACS (leave-one-domain-out) | average accuracy: 75.14% -> 76.95% | Improved |
| add MixStyle on two source domains (1 new domain) (ablation) | ResNet-18 ERM · Office-Home (leave-one-domain-out) | average accuracy: 61.91% -> 62.01% | No clear effect |
| add MixStyle on all three source domains (ablation) | ResNet-18 ERM · Office-Home target Product | accuracy: 71.15% -> 71.75% | No clear effect |
| add MixStyle on all three source domains (ablation) | ResNet-18 ERM · Office-Home target Real-World | accuracy: 74.74% -> 74.74% | No clear effect |
| add MixStyle on all three source domains (ablation) | ResNet-18 ERM · Digits-DG target MNIST-M | accuracy: 37.00% -> 63.65% | Improved |
| add MixStyle on all three source domains (ablation) | ResNet-18 ERM · PACS target Art | accuracy: 72.63% -> 80.12% | Improved |
| add OT-distance domain alignment on top of MixStyle (ablation) | ResNet-18 with MixStyle · Digits-DG, PACS, Office-Home generalization error | domain generalization error | Improved |
| add OT-distance alignment to ERM (ablation) | ResNet-18 ERM · PACS target Photo | domain generalization error | Hurt |
Closest papers by meaning
- Failure Modes of Domain Generalization Algorithms (CVPR 2022 · similarity 0.93)
- Exploiting Domain-Specific Features to Enhance Domain Generalization (NeurIPS 2021 · similarity 0.92)
- Robust Domain Generalization under Divergent Marginal and Conditional Distributions (NeurIPS 2026 · similarity 0.90)
- Unknown Domain Inconsistency Minimization for Domain Generalization (ICLR 2024 · similarity 0.89)
- One-Step Generalization Ratio Guided Optimization for Domain Generalization (ICML 2025 · similarity 0.89)
- Rethinking Multi-domain Generalization with A General Learning Objective (CVPR 2024 · 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.
