AI Paper Atlas

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

Calibrating Decision Robustness via Inverse Conformal Risk Control

Wenbin Zhou, Shixiang Zhu

ICML 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
increase calibration samples n from 10 to 30 (ablation)CREME · Linear Programming (|Lambda|=10)Pareto frontier gap: 0.10 ± 0.03 -> 0.08 ± 0.01Improved
increase calibration samples n from 10 to 30 (ablation)CREME · Newsvendor (|Lambda|=10)Pareto frontier gap: 0.19 ± 0.08 -> 0.13 ± 0.04Improved
increase calibration samples n from 10 to 30 (ablation)CREME · Portfolio Optimization (|Lambda|=10)Pareto frontier gap: 0.16 ± 0.07 -> 0.09 ± 0.02Improved
increase calibration samples n from 10 to 30 (ablation)CREME · Shortest Path (|Lambda|=10)Pareto frontier gap: 0.11 ± 0.03 -> 0.06 ± 0.01Improved
increase robustness index set size from 10 to 30 at n=30 (ablation)CREME · Linear ProgrammingPareto frontier gap: 0.08 ± 0.01 -> 0.06 ± 0.02Improved
increase robustness index set size from 10 to 30 at n=30 (ablation)CREME · Shortest PathPareto frontier gap: 0.06 ± 0.01 -> 0.04 ± 0.01Improved
increase robustness index set size from 10 to 30 at n=30 (ablation)CREME · NewsvendorPareto frontier gap: 0.13 ± 0.04 -> 0.11 ± 0.04Improved
increase robustness index set size from 10 to 30 at n=30 (ablation)CREME · Linear Programmingruntime (s): 1.00 ± 0.04 -> 2.98 ± 0.13Hurt
increase robustness index set size from 10 to 30 at n=30 (ablation)CREME · Portfolio Optimizationruntime (s): 1.50 ± 0.13 -> 4.49 ± 0.44Hurt
replace Monte Carlo average loss with conformal CREME estimatorMonte Carlo estimator · four optimization problemsvalidity vs accuracyMixed
select robustness level via CREME frontier instead of 95% quantile or conformal heuristicsfixed-coverage baselines · four optimization problemsmiscoverage-regret tradeoffImproved
add data-splitting recalibration after post-hoc selection (ablation)CREME · four optimization problemsvalidity of reported riskImproved

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