Research map / Label-efficient and robust learning / OOD and anomaly detection
Invariant Learning via Probability of Sufficient and Necessary Causes
Mengyue Yang, Zhen Fang, Yonggang Zhang, Yali Du, Furui Liu, Jean-Francois Ton, Jianhong Wang, Jun 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 |
|---|---|---|---|
| add PNS-based CaSN risk to ERM-style training (CaSN base vs ERM) | CaSN · PACS | average accuracy: 85.5 -> 86.0 | Improved |
| add PNS risk (CaSN base vs ERM) on worst domain | CaSN · PACS | worst-domain accuracy: 79.3 -> 80.2 | Improved |
| add MMD loss to CaSN (ablation) | CaSN · PACS | worst-domain accuracy: 80.2 -> 81.4 | Improved |
| add IRM loss to CaSN (ablation) | CaSN · PACS | average accuracy: 86.0 -> 83.5 | Hurt |
| CaSN base vs ERM on VLCS | CaSN · VLCS | average accuracy: 77.5 -> 77.2 | No clear effect |
| add IRM loss to CaSN (ablation) | CaSN · VLCS | average accuracy: 77.2 -> 78.2 | Improved |
| add MMD loss to CaSN (ablation) | CaSN · VLCS | worst-domain accuracy: 64.8 -> 65.9 | Improved |
| CaSN vs ERM on ColoredMNIST | CaSN · ColoredMNIST | average accuracy: 51.5 -> 52.2 | Improved |
| remove Monotonicity evaluator (CaSN(-m)) (ablation) | CaSN · synthetic data | distance correlation with SN | Hurt |
| larger delta (1.1 vs 0.1) (ablation) | CaSN · synthetic data | distance correlation with SN: 0.90 | Improved |
| stronger spurious degree s=0.7 vs s=0.1 | CaSN · synthetic data | distance correlation with spurious factor: 0.13 -> 0.37 | Hurt |
Closest papers by meaning
- Towards a Theoretical Framework of Out-of-Distribution Generalization (NeurIPS 2021 · similarity 0.91)
- Out-of-Distribution Generalization With Causal Invariant Transformations (CVPR 2022 · similarity 0.91)
- Diagnosing and Rectifying Fake OOD Invariance: A Restructured Causal Approach (AAAI 2024 · similarity 0.88)
- Transformation-Invariant Learning and Theoretical Guarantees for OOD Generalization (NeurIPS 2024 · similarity 0.86)
- On the Out-of-Distribution Generalization of Self-Supervised Learning (ICML 2025 · similarity 0.86)
- Out-of-Variable Generalisation for Discriminative Models (ICLR 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.
