Research map / Label-efficient and robust learning / Recognition and long-tail
ZIN: When and How to Learn Invariance Without Environment Partition?
LIN Yong, Shengyu Zhu, Lu Tan, Peng Cui
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 auxiliary information to jointly infer environments and learn invariant representations | ZIN · Temporal synthetic (p_s(t)=(0.999,0.7), p_v=0.9) | Test Worst accuracy: 57.31 -> 85.36 | Improved |
| Use auxiliary information to jointly infer environments and learn invariant representations | ZIN · House Price Prediction | Test Mean MSE: 0.9625 -> 0.3339 | Improved |
| Use seven rather than one auxiliary annotation (ablation) | ZIN on CelebA · CelebA | Test Worst accuracy: 47.62±0.98 -> 67.27±1.15 | Improved |
| Use location rather than climate variables as auxiliary information (ablation) | ZIN on Landcover · Landcover | OOD Test accuracy: 62.50 -> 66.06 | Improved |
| Add auxiliary information Z to EIIL (ablation) | EIIL on House Price Prediction · House Price Prediction | Test Mean MSE: 0.9625 -> 0.9701 | Hurt |
| Add auxiliary information Z to HRM (ablation) | HRM on House Price Prediction · House Price Prediction | Test Mean MSE: 0.4621 -> 0.4221 | Improved |
| Use location instead of climate variables as auxiliary information (ablation) | ZIN on Landcover · Landcover | OOD Test accuracy: 62.50 -> 66.06 | Improved |
| Vary the number of inferred environments K (ablation) | ZIN · Experiments (settings summarized in text) | Performance | No clear effect |
Closest papers by meaning
- Provably Invariant Learning without Domain Information (ICML 2023 · similarity 0.91)
- When Does Group Invariant Learning Survive Spurious Correlations? (NeurIPS 2022 · similarity 0.87)
- Unsupervised Representation Learning - an Invariant Risk Minimization Perspective (ICLR 2026 · similarity 0.87)
- The Implicit Bias of Heterogeneity towards Invariance: A Study of Multi-Environment Matrix Sensing (NeurIPS 2024 · similarity 0.86)
- Sufficient Invariant Learning for Distribution Shift (CVPR 2025 · similarity 0.85)
- Diagnosing and Rectifying Fake OOD Invariance: A Restructured Causal Approach (AAAI 2024 · similarity 0.85)
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.
