Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Bayesian Invariant Risk Minimization
Check 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 |
|---|---|---|---|
| estimate the IRM penalty over classifier posterior distributions rather than a single classifier | IRM · Synthetic dataset, 0.5K training samples | test MSE: 13.75 -> 3.86 | Helped |
| estimate the IRM penalty over classifier posterior distributions rather than a single classifier | IRM · CMNIST, 5K training samples | test accuracy: 32.55 -> 60.01 | Helped |
| estimate the IRM penalty over classifier posterior distributions rather than a single classifier | IRM · ColoredObject | test accuracy: 73.5±1.5 -> 78.1±0.6 | Helped |
| estimate the IRM penalty over classifier posterior distributions rather than a single classifier | IRM · CifarMnist | test accuracy: 52.3±0.9 -> 59.3±2.3 | Helped |
| reduce CMNIST training samples from 50K to 5K | IRMv1 · CMNIST | test accuracy: 67.45 -> 26.19 | Hurt |
| reduce CMNIST training samples from 50K to 5K | BIRM · CMNIST | test accuracy: 69.97 -> 60.01 | Hurt |
| increase IRMv1 L2 weight decay from 1 × 10−3 to 1 × 10−2 | IRMv1 · CMNIST, 10K training samples | test accuracy: 35.60 -> 49.78 | Helped |
| increase IRMv1 L2 weight decay from 1 × 10−2 to 1 × 10−1 | IRMv1 · CMNIST, 10K training samples | test accuracy: 49.78 -> 50.44 | No clear effect |
| add early stopping with weight decay 1 × 10−3 | IRMv1 · CMNIST, 10K training samples | test accuracy: 35.60 -> 41.92 | Helped |
Closest papers by meaning
- The Risks of Invariant Risk Minimization (ICLR 2021 · similarity 0.92)
- Invariant Risk Minimization Is A Total Variation Model (ICML 2024 · similarity 0.90)
- On the Asymptotic Distribution of the Minimum Empirical Risk (ICML 2024 · similarity 0.85)
- Sufficient Invariant Learning for Distribution Shift (CVPR 2025 · similarity 0.84)
- Out-of-distribution Generalization for Total Variation based Invariant Risk Minimization (ICLR 2025 · similarity 0.83)
- On the Efficiency of ERM in Feature Learning (NeurIPS 2024 · similarity 0.83)
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.