Research map / Label-efficient and robust learning / Semi-supervised learning
MEC: Machine-Learning-Assisted Generalized Entropy Calibration for Semi-Supervised Mean Estimation
Se Yoon Lee, Jae Kwang Kim
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 |
|---|---|---|---|
| calibrate labeled weights with Bregman projection on predictor basis (MEC) instead of CF-PPI | MEC · Synthetic semi-supervised mean estimation (N=1000, d=10), f in [0.1,0.5] | CI width ratio | Improved |
| use cross-fitting (out-of-fold predictor) instead of single-fit vanilla PPI | CF-PPI · Synthetic semi-supervised mean estimation (N=1000, d=10) | 95% CI coverage | Improved |
| use CF-PPI with KRR or kNN at small label fraction | CF-PPI · Synthetic semi-supervised mean estimation (N=1000, d=10), f=0.1 | CI width ratio | Hurt |
| use a strong predictor (KRR or RF) rather than weak FNN or kNN (ablation) | MEC · Synthetic semi-supervised mean estimation (N=1000, d=10) | CI width ratio | Improved |
| change generator of the Bregman divergence (other than quadratic) (ablation) | MEC · Synthetic semi-supervised mean estimation (N=1000, d=10) | coverage and width ratio | No clear effect |
Closest papers by meaning
- Prediction-Powered Semi-Supervised Learning with Online Power Tuning (NeurIPS 2025 · similarity 0.87)
- Regression for the Mean: Auto-Evaluation and Inference with Few Labels through Post-hoc Regression (ICML 2025 · similarity 0.86)
- Reassessing How to Compare and Improve the Calibration of Machine Learning Models (ICLR 2025 · similarity 0.85)
- Anytime-valid, Bayes-assisted, Prediction-Powered Inference (NeurIPS 2025 · similarity 0.85)
- The Minimax Rate of Second-Order Calibration (NeurIPS 2026 · similarity 0.84)
- Semi-Supervised Regression with Heteroscedastic Pseudo-Labels (NeurIPS 2025 · similarity 0.84)
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.
