AI Paper Atlas

Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes

Bayesian Post Training Enhancement of Regression Models with Calibrated Rankings

Kevin Tirta Wijaya, Bing Xu Hu, Hans-peter Seidel, Wojciech Matusik, Vahid Babaei

ICLR 2026 · Probabilistic and scientific modeling › Uncertainty and Gaussian processes · approach: Nonparametric and PAC-Bayes Bayesian learning

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 RankRefine++ instead of RankRefine with LLM rankerranker fusion · VDss, ChatGPT5 rankerMAE ratio: 0.952 -> 0.853Improved
use RankRefine++ instead of Projection with ChatGPT5 on FreeSolvranker fusion · FreeSolvMAE ratio: 1.119 -> 0.997Improved
use RankRefine++ instead of RankRefine on FreeSolv with ChatGPT5ranker fusion · FreeSolvMAE ratio: 1.002 -> 0.997No clear effect
use RankRefine++ instead of RankRefine on FreeSolv with Claude4ranker fusion · FreeSolvMAE ratio: 1.070 -> 0.977Improved
use RankRefine++ instead of RankRefine, median over datasetsranker fusion · 12 datasets, 65% ranker, k=30median MAE reduction: 9.78% -> 19.33%Improved
add temperature scaling to MLE (ablation)RankRefine++ · Clearance HepatocyteMAE ratioMixed
add accuracy-aware soft gating on top of temperature scaling (ablation)RankRefine++ · Clearance HepatocyteMAE ratioImproved
use MAP Gaussian prior instead of MLE without temperature (ablation)RankRefine++ · Clearance Hepatocyte large kMAE ratioImproved
apply RankRefine++ to RF and MLP regressors on non-molecular datasetsRankRefine++ · 3 tabular datasetsMAE ratioImproved
RankRefine versus RankRefine++ under Gaussian assumptionsRankRefine++ · 5 ADMET tasksmean abs difference: 0.028 ± 0.028No 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