AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

Fast Conformal Prediction Using Conditional Interquantile Intervals

Naixin Guo, Rui Luo, Zhixin Zhou

AAAI 2026 · Learning theory and algorithms › Prediction and decision losses · approach: Conformal prediction sets · 5 citations (OpenAlex)

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 changedWhereMeasuredEffect
use CIR instead of CQR on skewed dataconformal regression with deep quantile model · Synthetic (|I|=5000, skew 2.7)interval width: 5.345 -> 3.776Improved
add width-based selection rule (CIR+ vs CIR) (ablation)CIR · Synthetic (|I|=5000, skew 2.7)interval width: 3.776 -> 3.643Improved
add width-based selection rule (CIR+ vs CIR) (ablation)CIR · Synthetic (|I|=5000, skew 2.7)time: 0.663 -> 0.862Hurt
replace histogram construction of CHR with interquantile intervals (CIR)CHR · Synthetic (|I|=5000, skew 2.7)time: 103.637 -> 0.663Improved
replace histogram construction of CHR with interquantile intervals (CIR)CHR · Synthetic (|I|=5000, skew 2.7)interval width: 3.389 -> 3.776Hurt
increase sample sizeCHR · Synthetic (skew 2.7)time: 36.597 (|I|=500) -> 103.637 (|I|=5000)Hurt
increase distribution symmetry (skew 2.70 to 0.66)CIR and CQR vs CHR · Synthetic (|I|=5000)interval width: CQR minus CIR+ gap: 5.345 vs 3.643 at skew 2.70; 5.950 vs 5.715 at skew 0.66Mixed
use CIR with small calibration/training dataCIR/CIR+ · Synthetic (|I|=500)interval width: 3.930 (|I|=500) vs 3.776 (|I|=5000)Hurt
CIR+ vs CIR marginal coverage tie handling (ablation)CIR · Synthetic (|I|=5000, skew 2.7)marginal coverage: 0.905 -> 0.900Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy