AI Paper Atlas

Research map / Label-efficient and robust learning / Noisy labels

Learning with Admissibility: Robust Fuzzy Hashing for Cross-Modal Retrieval with Noisy Labels

Xincheng Sun, Ruitao Pu, Guangsi Shi, Zhenwen Ren, Peng Hu, Yuan Sun

ICML 2026 · Label-efficient and robust learning › Noisy labels · approach: Noisy cross-modal hashing and multi-label

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 warm-up stage (ablation)RFCMH · XMedia noise 0.6summed mAP: 355.7 -> 359.3Improved
add warm-up stage (ablation)RFCMH · INRIA-Websearch noise 0.6summed mAP: 187.1 -> 201.0Improved
add L_FAR (fuzzy admissibility supervision) (ablation)RFCMH · INRIA-Websearch noise 0.6summed mAP: 168.0 -> 201.0Improved
add L_FAR (ablation)RFCMH · XMediaNet noise 0.6summed mAP: 184.1 -> 196.8Improved
add DGSA (ablation)RFCMH · XMediaNet noise 0.6summed mAP: 37.9 -> 196.8Improved
add DGSA (ablation)RFCMH · XMedia noise 0.6summed mAP: 195.4 -> 359.3Improved
add necessity measure in FAR (ablation)RFCMH · XMediaNet noise 0.6summed mAP: 194.8 -> 196.8Improved
add necessity measure in FAR (ablation)RFCMH · XMedia noise 0.6summed mAP: 355.7 -> 359.3No clear effect
add FAR under clean labels (ablation)RFCMH · INRIA-Websearch I2T, 128 bits, cleanmAP: 49.5 -> 55.8Improved
add DGSA under clean labels (ablation)RFCMH · XMediaNet I2T, 16 bits, cleanmAP: 0.9 -> 43.7Improved
increase alpha beyond about 0.1 (ablation)RFCMH · INRIA-Websearch 128 bitsmAPHurt
use sample selection (hard separation) baseline instead of fuzzy supervisionNRCH · XMediaNet (200 classes), noise 0.6I2T mAP at 128 bits: 54.2 (RFCMH) vs 19.7 (NRCH)Hurt

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