Research map / Probabilistic and scientific modeling / Molecules and proteins
No Free Lunch: Non-Asymptotic Analysis of Prediction-Powered Inference
Pranav Mani, Peng Xu, Zachary Chase Lipton, Michael Oberst
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 single-sample PPI++ (no sample splitting) instead of classical estimator for interval construction | PPI++ confidence intervals · AlphaFold, n = 20 | coverage: realized coverage 87% at n = 20 | Hurt |
| add cross-fitting (sample splitting) to PPI++ (ablation) | PPI++ · AlphaFold | interval coverage | Improved |
| lower pseudo-label correlation below roughly 1/sqrt(n) | PPI++ vs classical estimator · Gaussian labels | variance: improvement requires correlation above 1/sqrt(n - 2) | Hurt |
| use cross-fit instead of single-sample PPI++ | PPI++ · Gaussian labels | correlation threshold for improvement: (n - 2)^-1/2 -> (n/2 - 2)^-1/2 | Hurt |
| increase labeled sample size n | PPI++ with imperfect pseudo-labels · AlphaFold | MSE difference vs classical | Improved |
Closest papers by meaning
- Regression for the Mean: Auto-Evaluation and Inference with Few Labels through Post-hoc Regression (ICML 2025 · similarity 0.92)
- Prediction-Powered Semi-Supervised Learning with Online Power Tuning (NeurIPS 2025 · similarity 0.89)
- Anytime-valid, Bayes-assisted, Prediction-Powered Inference (NeurIPS 2025 · similarity 0.89)
- Prediction-Powered Adaptive Shrinkage Estimation (ICML 2025 · similarity 0.89)
- FAB-PPI: Frequentist, Assisted by Bayes, Prediction-Powered Inference (ICML 2025 · similarity 0.89)
- Stratified Prediction-Powered Inference for Effective Hybrid Evaluation of Language Models (NeurIPS 2024 · similarity 0.87)
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.
