Research map / Learning theory and algorithms / Sample complexity
Empirical or Invariant Risk Minimization? A Sample Complexity Perspective
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 IRM instead of ERM | 2-layer MLP · CS-CMNIST, test environment, 60000 samples | test error: 8.15 ± 0.37 -> 9.62 ± 0.67 | No clear effect |
| Use IRM instead of ERM | 2-layer MLP · AC-CMNIST, test environment, 60000 samples | test error: 71.96 ± 0.55 -> 32.78 ± 2.70 | Helped |
| Use IRM instead of ERM | 2-layer MLP · CF-CMNIST, test environment, 60000 samples | test error: 67.62 ± 0.31 -> 44.92 ± 0.84 | Helped |
| Use IRM instead of ERM | 2-layer MLP · HB-CMNIST, test environment, 60000 samples | test error: 59.70 ± 0.72 -> 32.80 ± 0.55 | Helped |
| Use IRM instead of ERM | linear model · CS-regression, n=10, 2000 samples | model estimation error: 0.018 ± 0.002 -> 0.035 ± 0.006 | No clear effect |
| Use IRM instead of ERM | linear model · CF-regression, n=10, 2000 samples | model estimation error: 0.07 ± 0.01 -> 0.07 ± 0.01 | No clear effect |
| Use IRM instead of ERM | linear model · AC-regression, n=10, 2000 samples | model estimation error: 0.77 ± 0.08 -> 0.54 ± 0.04 | Helped |
| Use IRM instead of ERM | linear model · HB-regression, n=10, 2000 samples | model estimation error: 0.74 ± 0.08 -> 0.58 ± 0.05 | Helped |
| Use IRM instead of ERM | linear predictor under confounder or anti-causal shifts · Linear and polynomial generative models | distance to OOD solution: O(1/ε^2); O(√ε) | Helped |
| Use IRM instead of ERM | predictor under covariate shift · Covariate shift | sample complexity | No clear effect |
Closest papers by meaning
- The Risks of Invariant Risk Minimization (ICLR 2021 · similarity 0.91)
- Bayesian Invariant Risk Minimization (CVPR 2022 · similarity 0.90)
- On the Asymptotic Distribution of the Minimum Empirical Risk (ICML 2024 · similarity 0.89)
- Invariant Risk Minimization Is A Total Variation Model (ICML 2024 · similarity 0.88)
- Out-of-distribution Generalization for Total Variation based Invariant Risk Minimization (ICLR 2025 · similarity 0.86)
- On the Efficiency of ERM in Feature Learning (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.