AI Paper Atlas

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

Cost-Sensitive Conformal Training with Provably Controllable Learning Bounds

Xuesong Jia, Yuanjie Shi, Ziquan Liu, Yi Xu, Yan Yan

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
train with RWCE instead of ConfTrResNet · Caltech-101 (HPS)APSS: 1.26 -> 0.96Improved
train with RWCE instead of ConfTrResNet · Caltech-101 (APS)APSS: 4.25 -> 1.33Improved
train with RWCE instead of ConfTrDenseNet · Caltech-101 (APS)APSS: 9.69 -> 1.27Improved
train with RWCE instead of CUTResNet · CIFAR-100 (HPS)APSS: 2.91 -> 2.68Improved
train with RWCE instead of CUTDenseNet · CIFAR-100 (HPS)APSS: 2.07 -> 2.04Improved
train with RWCE instead of CUTDenseNet · CIFAR-100 (APS)APSS: 2.19 -> 2.335Hurt
train with RWCE instead of ConfTrDenseNet · iNaturalist (HPS)APSS: 72.25 -> 67.17Improved
train with RWCE instead of CUTDenseNet · iNaturalist (APS)APSS: 87.27 -> 75.65Improved
train with RWCE instead of strongest baseline, averaged over configurationsRWCE · three datasetsAPSS: 21.38% reductionImproved
train with ConfTr instead of CEResNet · CIFAR-100 (HPS)APSS: 3.39 -> 3.98Hurt
train with ConfTr instead of CEDenseNet · Caltech-101 (APS)APSS: 8.92 -> 9.69Hurt

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