Research map / Label-efficient and robust learning / Domain adaptation and generalization
Invariant and Transportable Representations for Anti-Causal Domain Shifts
Yibo Jiang, Victor Veitch
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 |
|---|---|---|---|
| ACTIR instead of ERM | ACTIR · Synthetic | test accuracy: 9.95 -> 74.77 | Improved |
| fine-tune ACTIR on 10 labeled examples | ACTIR · Synthetic | accuracy: 74.77 -> 89.28 | Improved |
| fine-tune IRM on 10 labeled examples | IRM · Synthetic | accuracy: 74.91 -> 74.27 | No clear effect |
| ACTIR instead of IRM | ACTIR · Color MNIST | test accuracy: 59.97 -> 70.30 | Improved |
| fine-tune ACTIR on 10 labeled examples | ACTIR · Color MNIST | accuracy: 70.30 -> 85.25 | Improved |
| ACTIR instead of IRM | ACTIR · Camelyon17 | test accuracy: 71.59 -> 77.73 | Improved |
| ACTIR instead of IRM | ACTIR · PACS Art | accuracy: 78.69 -> 82.55 | Improved |
| ACTIR instead of IRM | ACTIR · PACS Sketch | accuracy: 65.61 -> 62.14 | Hurt |
| ACTIR instead of ERM | ACTIR · VLCS Caltech | accuracy: 95.29 -> 95.15 | No clear effect |
| ACTIR instead of MAML | ACTIR · VLCS SUN09 | accuracy: 70.84 -> 72.31 | Improved |
| ACTIR instead of MAML | ACTIR · VLCS Caltech | accuracy: 96.75 -> 95.15 | Hurt |
| data violates anti-causal structure: ACTIR instead of IRM | ACTIR · Synthetic (non anti-causal) | test accuracy: 69.61 -> 43.51 | Hurt |
Closest papers by meaning
- Measure-Theoretic Anti-Causal Representation Learning (NeurIPS 2025 · similarity 0.90)
- Learning Linear Causal Representations from General Environments: Identifiability and Intrinsic Ambiguity (NeurIPS 2024 · similarity 0.89)
- Unifying Causal Representation Learning with the Invariance Principle (ICLR 2025 · similarity 0.88)
- Diagnosing and Rectifying Fake OOD Invariance: A Restructured Causal Approach (AAAI 2024 · similarity 0.88)
- Identifying General Mechanism Shifts in Linear Causal Representations (NeurIPS 2024 · similarity 0.88)
- A Sparsity Principle for Partially Observable Causal Representation Learning (ICML 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.
