AI Paper Atlas

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

CARE: Adaptive Calibration for Reliable Recommendations

Nitin Bisht, Huan Huo, Xiuwen Gong, Guandong Xu

ICML 2026 · Learning theory and algorithms › Prediction and decision losses · approach: Probabilistic calibration

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
wrap NeuMF with CARE adaptive calibrationNeuMF · Book-CrossingMRR / Recall@10: MRR 0.246 -> 0.289; Recall 0.502 -> 0.557Improved
wrap FMLP-Rec with CAREFMLP-Rec · Book-CrossingMRR / Recall@10: 0.335 -> 0.357; 0.599 -> 0.628Improved
wrap SASRec with CARESASRec · Last.fmRecall@10: 0.766 -> 0.799Improved
wrap CASER with CARECASER · Last.fmNDCG@10: 0.367 -> 0.385Improved
use CARE-wrapped backbone instead of preference-aware Oracle4RecOracle4Rec · Book-CrossingRecall@10: 0.603 -> 0.628 (FMLP-Rec + CARE)Improved
use CARE instead of fixed-threshold Split ConformalSplit Conformal · Book-Crossing, alpha=0.10, SASReccoverage: 0.835 -> 0.908Improved
use CARE instead of EnbPI fixed-window calibrationEnbPI · Book-Crossing, alpha=0.10, SASRecset size: 47 -> 43Improved
use CARE instead of Online ConformalOnline Conformal · Book-Crossing, alpha=0.10, FMLP-Recset size: 45 -> 42Improved
use CARE instead of Split Conformal on NeuMFSplit Conformal · Book-Crossing, alpha=0.10, NeuMFset size: 44 -> 46Hurt
add CARE online calibration phase to backbone trainingSASRec · Book-Crossing training time (minutes)training time: 35.2 -> 36.5Hurt

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