Research map / Label-efficient and robust learning / Recognition and long-tail
When Does Group Invariant Learning Survive Spurious Correlations?
Yimeng Chen, Ruibin Xiong, Zhi-Ming Ma, Yanyan Lan
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 |
|---|---|---|---|
| replace EIIL groups and objective with SCILL (IRM penalty) | group invariant learning · PC-MNIST test, ID selection | accuracy: 50.63 -> 62.49 | Improved |
| replace EIIL with SCILL (REx), TEV selection | group invariant learning · PC-MNIST test | accuracy: 51.01 -> 67.33 | Improved |
| replace EIIL with SCILL (REx) | BERT base · MNLI-HANS test, ID selection | accuracy: 65.16 -> 69.73 | Improved |
| replace EIIL with SCILL (cMMD) | BERT base · MNLI-HANS test, ID selection | accuracy: 63.22 -> 69.15 | Improved |
| SCILL-REx with ID selection vs PoE and ConfReg | BERT base · MNLI-HANS | accuracy: 69.2 -> 69.73 | Improved |
| remove instance reweighting (SCILL uw) (ablation) | SCILL-IRM · PC-MNIST test, Oracle selection | accuracy: 67.46 -> 53.51 | Hurt |
| remove instance reweighting (SCILL uw) (ablation) | SCILL-IRM · MNLI-HANS test, ID selection | accuracy: 69.11 -> 64.30 | Hurt |
| use majority/minority groups instead of ground-truth strata (ablation) | IRM · PC-MNIST test, Oracle selection | accuracy: 70.19 -> 63.85 | Hurt |
| unbalance label weights (p_err 1 -> 2) (ablation) | SCILL-gt-IRM · PC-MNIST test, TEV selection | accuracy: 70.91 -> 56.20 | Hurt |
| unbalance label weights (p_err 1 -> 1.2) (ablation) | SCILL-gt-IRM · PC-MNIST test, TEV selection | accuracy: 70.91 -> 63.43 | Hurt |
| lower statistical split threshold from 10 to 5 (ablation) | SCILL-IRM · PC-MNIST test, ID selection | accuracy: 62.49 -> 55.50 | Hurt |
| raise threshold from 10 to 20 (ablation) | SCILL-IRM · PC-MNIST test, Oracle selection | accuracy: 67.46 -> 67.95 | No clear effect |
Closest papers by meaning
- Improving Group Robustness on Spurious Correlation Requires Preciser Group Inference (ICML 2024 · similarity 0.88)
- Sufficient Invariant Learning for Distribution Shift (CVPR 2025 · similarity 0.87)
- Let Samples Speak: Mitigating Spurious Correlation by Exploiting the Clusterness of Samples (CVPR 2025 · similarity 0.87)
- The Group Robustness is in the Details: Revisiting Finetuning under Spurious Correlations (NeurIPS 2024 · similarity 0.86)
- Unsupervised Concept Discovery Mitigates Spurious Correlations (ICML 2024 · similarity 0.85)
- Decompose-and-Compose: A Compositional Approach to Mitigating Spurious Correlation (CVPR 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.
