Research map / Label-efficient and robust learning / Recognition and long-tail
Causal Balancing for Domain Generalization
Xinyi Wang, Michael Saxon, Jiachen Li, Hongyang Zhang, Kun Zhang, William Yang Wang
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 |
|---|---|---|---|
| Use balanced mini-batch sampling with CORAL | CORAL · DomainBed average | accuracy: 67.5 -> 69.7 | Improved |
| Use balanced mini-batch sampling with CORAL | CORAL · ColoredMNIST | accuracy: 51.5 -> 66.6 | Improved |
| Use balanced mini-batch sampling with ERM | ERM · ColoredMNIST | accuracy: 51.5 -> 60.1 | Improved |
| Use balanced mini-batch sampling with IRM | IRM · ColoredMNIST | accuracy: 52.0 -> 59.2 | Improved |
| Use balanced mini-batch sampling with GroupDRO | GroupDRO · ColoredMNIST | accuracy: 52.1 -> 53.9 | Improved |
| Use balanced mini-batch sampling with ERM | ERM · RotatedMNIST | accuracy: 98.0 -> 97.7 | Hurt |
| Use balanced mini-batch sampling with IRM | IRM · VLCS | accuracy: 78.5 -> 76.5 | Hurt |
| Use balanced mini-batch sampling with CORAL | CORAL · PACS | accuracy: 86.2 -> 86.7 | Improved |
| Use balanced mini-batch sampling with CORAL | CORAL · Office-Home | accuracy: 68.7 -> 69.6 | Improved |
| Use balanced mini-batch sampling with ERM | ERM · TerraIncognita | accuracy: 46.1 -> 48.0 | Improved |
| Use balanced mini-batch sampling with CORAL | CORAL · DomainNet | accuracy: 41.5 -> 43.9 | Improved |
| Increase the fraction of balanced examples in a mini-batch (ablation) | ERM with oracle color matching · ColoredMNIST¹⁰ | accuracy | 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.
- SelfReg — SelfReg: Self-Supervised Contrastive Regularization for Domain Generalization (ICCV 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
- Lost Domain Generalization Is a Natural Consequence of Lack of Training Domains (AAAI 2024 · similarity 0.89)
- One-Step Generalization Ratio Guided Optimization for Domain Generalization (ICML 2025 · similarity 0.87)
- Balanced Direction from Multifarious Choices: Arithmetic Meta-Learning for Domain Generalization (CVPR 2025 · similarity 0.87)
- On the Out-of-Distribution Generalization of Self-Supervised Learning (ICML 2025 · similarity 0.87)
- Understanding Domain Generalization: A Noise Robustness Perspective (ICLR 2024 · similarity 0.87)
- Do causal predictors generalize better to new domains? (NeurIPS 2024 · similarity 0.86)
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.
