AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

Calibrated Information Bottleneck for Trusted Multi-modal Clustering

Shizhe Hu, Zhangwen Gou, Shuaiju Li, Jin Qin, Xiaoheng Jiang, Pei Lv, Mingliang Xu

ICLR 2026 · Vision-language and multimodal › Cross-modal learning · approach: Medical multimodal fusion

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 calibration loss LcaliH on top of LIB + LcluH (ablation)CLIB · Caltech-3VACC (%): 60.6 -> 67.8Improved
add LcaliH on top of LIB + LcluH (ablation)CLIB · Caltech-3VECE (%): 25.5 -> 11.6Improved
add reconstruction loss Lre (ablation)CLIB · Caltech-3VACC (%): 60.6 -> 64.6Improved
add consistency loss Lcon (ablation)CLIB · Caltech-3VECE (%): 25.5 -> 12.4Improved
add Lcon (ablation)CLIB · Caltech-3VACC (%): 60.6 -> 61.1No clear effect
add Lcon (ablation)CLIB · IAPRACC (%): 39.1 -> 37.8Hurt
add warm-up stage for IB before calibration (ablation)CLIB · Caltech-3VACC (%): 61.7 -> 77.8Improved
stop gradient from calibration heads to IB (ablation)CLIB · Caltech-3VACC (%): 65.9 -> 77.8Improved
use dynamic quality-score screening (S2) instead of max-probability selection (ablation)CLIB · Caltech-3VACC (%): 70.6 -> 77.8Improved
use screening strategy S2 instead of S1 (ablation)CLIB · Caltech-3VACC (%): 65.6 -> 77.8Improved
use screening strategy S2 instead of S3 (ablation)CLIB · MIRFlickrACC (%): 43.1 -> 55.4Improved
continue calibration training from epoch 101 to 200 on difficult samplesCLIB · Caltech-3V (hard 20% samples)ACC (%): 18.9 -> 46.7Improved

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