AI Paper Atlas

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

Label What Matters: Modality-Balanced and Difficulty-Aware Multimodal Active Learning

Yuqiao Zeng, Xu Wang, Tengfei Liang, Yiqing Hao, Yi Jin, Hui Yu

CVPR 2026 · Vision-language and multimodal › Cross-modal learning · approach: Multimodal fusion with missing modalities · 1 citations (OpenAlex)

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 RL-MBA instead of BMMAL (strongest baseline)multimodal active learning · Food101 3,000 labelsTop-1 accuracy: 0.8609 -> 0.8650Improved
use RL-MBA instead of BMMALmultimodal active learning · KineticsSound 3,000 labelsTop-1 accuracy: 0.4745 -> 0.4841Improved
use RL-MBA instead of BMMALmultimodal active learning · VGGSound 3,000 labelsTop-1 accuracy: 0.2053 -> 0.2223Improved
use active learning strategies instead of random selectionbaseline AL methods · VGGSound 3,000 labelsTop-1 accuracy: Random 0.2173 vs BMMAL 0.2053, BADGE 0.2023Hurt
add AMCB only to BMMAL baseline (ablation)RL-MBA · Food101 / KineticsSound / VGGSoundTop-1 accuracy: 0.8609 -> 0.8621; 0.4745 -> 0.4771; 0.2053 -> 0.2059Improved
add EFDA only to BMMAL baseline (ablation)RL-MBA · VGGSoundTop-1 accuracy: 0.2053 -> 0.2177Improved
combine AMCB and EFDA (full RL-MBA) (ablation)RL-MBA · VGGSoundTop-1 accuracy: 0.2177 -> 0.2223Improved
use Relative reward instead of Absolute or Incremental reward (ablation)RL-MBA policy · Food101Top-1 accuracyImproved
use RL policy selection instead of heavy selection heuristicsRL-MBA · Food101 per-round timeround total time (s): 1151.39 -> 884.39 (vs BMMAL)Improved
use RL-MBA instead of BMMAL on video-centric classesKineticsSound class-wise · KineticsSound 5,000 labelsper-class accuracyMixed

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