Research map / Probabilistic and scientific modeling / Molecules and proteins
Assumption-lean and data-adaptive post-prediction inference
Jiacheng Miao, Xinran Miao, Yixuan Wu, Jiwei Zhao, Qiongshi Lu
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 data-adaptive weight instead of fixed weight 1 (PPI) (ablation) | PSPA · UK Biobank DXA-BMD, biological sex | SE ratio vs classical: 1.363 -> 0.892 | Improved |
| use data-adaptive weight instead of EIF* weight (ablation) | PSPA · UK Biobank DXA-BMD, age | SE ratio vs classical: 1.391 -> 0.907 | Improved |
| use per-coordinate weights instead of one shared scalar (PPI++) (ablation) | PSPA · UK Biobank DXA-BMD, alcohol | SE ratio vs classical: 0.926 -> 0.908 | Improved |
| use per-coordinate weights instead of one shared scalar (PPI++) (ablation) | PSPA · UK Biobank DXA-BMD, biological sex | SE ratio vs classical: 0.920 -> 0.892 | Improved |
| add PPI fixed-weight prediction augmentation to classical estimator | PPI · UK Biobank DXA-BMD, physical activity | SE ratio vs classical: 1 -> 1.560 | Hurt |
| add EIF*-based augmentation to classical estimator | EIF*-based · UK Biobank DXA-BMD, smoking | SE ratio vs classical: 1 -> 1.423 | Hurt |
| add PSPA augmentation to classical estimator | PSPA · UK Biobank DXA-BMD, SB | SE ratio vs classical: 1 -> 0.914 | Improved |
| increase unlabeled sample size and imputation accuracy | PSPA · Simulation, linear and logistic regression | CI width ratio vs classical | Improved |
| use PPI or EIF* with poor ML accuracy or limited unlabeled data | PPI and EIF*-based · Simulation, ML-predicted labels | CI width ratio vs classical | Hurt |
| use imputation-based inference with ML-predicted covariates | imputation-based method · Simulation, ML-predicted covariates | CI coverage | Hurt |
| use PSPA with ML-predicted covariates instead of classical | PSPA · Simulation, ML-predicted covariates | CI width ratio vs classical | Improved |
| use PSPA instead of PPI/EIF*/PPI++ for sex-biased genes | PSPA · GTEx sex-biased gene expression, 44 tissues | number of sex-biased genes identified | Improved |
Closest papers by meaning
- Task-Agnostic Machine-Learning-Assisted Inference (NeurIPS 2024 · similarity 0.89)
- Prediction-Augmented Trees for Reliable Statistical Inference (NeurIPS 2026 · similarity 0.85)
- Post-ADC Inference: Valid Inference After Active Data Collection (NeurIPS 2026 · similarity 0.85)
- Anytime-valid, Bayes-assisted, Prediction-Powered Inference (NeurIPS 2025 · similarity 0.84)
- Reproduce, Replicate, Reevaluate. The Long but Safe Way to Extend Machine Learning Methods (AAAI 2024 · similarity 0.84)
- Active Statistical Inference (ICML 2024 · 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.
