Research map / Learning theory and algorithms / Prediction and decision losses
Conformalized Quantile Regression
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 conformalized quantile regression instead of standard split conformal | CQR versus split conformal · Simulated heteroscedastic data with outliers | Avg. Length: 2.91 -> 1.99 | Helped |
| use conformalized quantile regression instead of locally adaptive conformal | CQR versus locally adaptive conformal · Simulated heteroscedastic data with outliers | Avg. Length: 2.86 -> 1.99 | Helped |
| use CQR with quantile random forests instead of random-forest local conformal prediction | 11-dataset comparison · 11 regression datasets (average) | Avg. Length: 1.82 -> 1.41 | Helped |
| use CQR with quantile neural networks instead of neural-network local conformal prediction | 11-dataset comparison · 11 regression datasets (average) | Avg. Length: 1.81 -> 1.40 | Helped |
| use quantile neural networks without conformalization | Quantile Neural Net · 11 regression datasets (average) | Avg. Coverage: 90.05 -> 88.51 | Hurt |
| add quantile crossing post-processing (ablation) | CQR Neural Net · 11 regression datasets (average) | Avg. Length: 1.40 -> 1.35 | Helped |
| add quantile crossing post-processing (ablation) | Quantile Neural Net · 11 regression datasets (average) | Avg. Length: 1.49 -> 1.41 | Helped |
| use two-tailed asymmetric conformalization (ablation) | CQR Neural Net · 11 regression datasets (average) | Avg. Length: 1.40 -> 1.58 | Hurt |
| use two-tailed asymmetric conformalization (ablation) | CQR Random Forests · 11 regression datasets (average) | Avg. Length: 1.41 -> 1.57 | Hurt |
| estimate a conditional median instead of a conditional mean for locally adaptive conformal prediction (ablation) | Locally adaptive conformal prediction · Simulated heteroscedastic data with outliers | Avg. Length: 2.86 -> 2.86 | No clear effect |
| tune the underlying quantile levels by cross-validation (ablation) | CQR · 11 regression datasets (average) | Avg. Length | Helped |
| use CQR instead of non-conformalized quantile regression | CQR Neural Net versus Quantile Neural Net · 11 regression datasets (average) | Avg. Coverage: 88.51 -> 90.05 | Helped |
Closest papers by meaning
- Rectifying Conformity Scores for Better Conditional Coverage (ICML 2025 · similarity 0.92)
- Co-optimization for Adaptive Conformal Prediction (NeurIPS 2026 · similarity 0.91)
- Probabilistic Conformal Prediction with Approximate Conditional Validity (ICLR 2025 · similarity 0.90)
- A Unified Comparative Study with Generalized Conformity Scores for Multi-Output Conformal Regression (ICML 2025 · similarity 0.88)
- Online conformal prediction with decaying step sizes (ICML 2024 · similarity 0.88)
- Conformal Thresholded Intervals for Efficient Regression (AAAI 2025 · similarity 0.88)
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.