AI Paper Atlas

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

Quantifying and Improving Adaptivity in Conformal Prediction Through Input Transformations

Sooyong Jang, Insup Lee

AAAI 2026 · Learning theory and algorithms › Prediction and decision losses · approach: Conformal prediction sets · 0 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
apply difficulty-binned group-conditional CP (O-LAC) instead of LACLAC · ImageNet val alpha=0.10 (avg of 9 classifiers)T-SS: 0.543 -> 0.737Improved
apply difficulty-binned group-conditional CP (O-LAC) instead of LACLAC · ImageNet val alpha=0.10 (avg of 9 classifiers)T-CV: 0.304 -> 0.050Improved
apply difficulty-binned group-conditional CP (O-SAPS) instead of SAPSSAPS · ImageNet val alpha=0.10T-CV: 0.148 -> 0.045Improved
apply difficulty-binned group-conditional CP (O-SAPS) instead of SAPSSAPS · ImageNet val alpha=0.10T-SS: 0.673 -> 0.720Improved
apply O-LAC vs APS on T-CVAPS · ImageNet val alpha=0.10T-CV: 0.065 -> 0.050No clear effect
apply O-LAC vs APS on T-SS for ResNet50APS · ImageNet val ResNet50 alpha=0.10T-SS: 0.210 -> 0.835Improved
use O-LAC at alpha=0.05 instead of LACLAC · ImageNet val alpha=0.05T-SS: 0.678 -> 0.743Improved
use more accurate base classifier (ViT-H-14 vs ResNet18)O-LAC · ImageNet val alpha=0.10T-SS: 0.926 -> 0.627Hurt
increase target miscoverage alpha to 0.15-0.20T-SS metric · ImageNet valT-SSHurt
bin by difficulty and run per-bin thresholds (group-conditional CP)O-LAC · ImageNet valaverage set sizeMixed
use O-CPA/O-PAC instead of baselines for interval width vs errorPAC · Visual acuity alpha=0.30, EfficientNet-V2-SR2: 0.89 -> 0.92Improved
use O-CPA instead of CP-A for interval width vs errorCP-A · Visual acuity alpha=0.30, EfficientNet-V2-SR2: 0.87 -> 0.87No clear effect

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