AI Paper Atlas

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

MOON2.0: Dynamic Modality-balanced Multimodal Representation Learning for E-commerce Product Understanding

Zhanheng Nie, Chenghan Fu, Daoze Zhang, Junxian Wu, Wanxian Guan, Pengjie Wang, Jian Xu, Bo Zheng

CVPR 2026 · Vision-language and multimodal › Cross-modal learning · approach: Contrastive vision-language alignment · 5 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
add Modality-driven MoE (vs plain MLP) (ablation)MOON2.0 · MBE2.0 q_i->c_mmR@10: 74.59 -> 91.08Improved
add Modality-driven MoE (vs plain MLP) (ablation)MOON2.0 · MBE2.0 product classificationaccuracy: 62.55 -> 68.08Improved
add Modality-driven MoE (vs plain MLP) (ablation)MOON2.0 · MBE2.0 attribute predictionaccuracy: 75.62 -> 84.29Improved
add Dual-level Alignment (ablation)MOON2.0 · MBE2.0 q_t->c_iR@10: 31.45 -> 73.12Improved
add Dual-level Alignment (ablation)MOON2.0 · MBE2.0 q_i->c_tR@10: 23.35 -> 64.91Improved
add Dual-level Alignment (ablation)MOON2.0 · MBE2.0 q_t->c_mmR@10: 37.99 -> 63.09Improved
add MLLM-based Image-text Co-augmentation (ablation)MOON2.0 · MBE2.0 q_i->c_mmR@10: 78.17 -> 91.08Improved
add MLLM-based Image-text Co-augmentation (ablation)MOON2.0 · MBE2.0 attribute predictionaccuracy: 77.77 -> 84.29Improved
add Dynamic Sample Filtering (ablation)MOON2.0 · MBE2.0 q_mm->c_mmR@10: 80.00 -> 94.21Improved
add Dynamic Sample Filtering (ablation)MOON2.0 · MBE2.0 attribute predictionaccuracy: 84.04 -> 84.29No clear effect
add Dynamic Sample Filtering (ablation)MOON2.0 · MBE2.0 product classificationaccuracy: 67.99 -> 68.08No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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