AI Paper Atlas

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

Calibration by Distribution Matching: Trainable Kernel Calibration Metrics

Charles Thomas Marx, Sofian Zalouk, Stefano Ermon

NeurIPS 2023 · Learning theory and algorithms › Prediction and decision losses · approach: Distributionally robust optimization

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
add MMD calibration regularizer to NLL trainingregression MLP · crimequantile calibration error: 0.220 -> 0.164Improved
add MMD calibration regularizer to NLL trainingregression MLP · superconductivityNLL: 3.375 -> 3.269Improved
add MMD calibration regularizer to NLL trainingregression MLP · blogNLL: 0.997 -> 0.957Improved
add MMD calibration regularizer to NLL trainingregression MLP · crimedecision calibration error: 0.151 -> 0.042Improved
add post-hoc quantile recalibration to NLL-trained modelregression MLP · crimeNLL: -0.716 -> -0.387Hurt
add post-hoc quantile recalibration to NLL + MMD modelregression MLP · crimequantile calibration error: 0.164 -> 0.105Improved
add post-hoc recalibration to NLL + MMD modelregression MLP · medical-expenditureNLL: 1.532 -> 1.780Hurt
add MMD regularizer before post-hoc recalibration (reduces sharpness penalty)regression MLP · fb-commentNLL after post-hoc: 0.734 -> 0.639Improved
add post-hoc recalibration to NLL-trained modelregression MLP · blogNLL: 0.997 -> 0.624Improved
replace cross-entropy with XE + MMD regularizerclassification MLP · breast-cancerECE: 0.194 -> 0.052Improved
replace cross-entropy with XE + MMD regularizerclassification MLP · heart-diseaseaccuracy: 55.904 -> 61.516Improved
use MMD instead of KDE regularizerclassification MLP · heart-diseaseaccuracy: 50.036 -> 61.516Improved

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