Research map / Vision-language and multimodal / Human and embodied modeling
M$^{3}$3D: A Multimodal, Multilingual and Multitask Dataset for Grounded Document-Level Information Extraction
Jiang Liu, Bobo Li, Xinran Yang, Na Yang, Hao Fei, Mingyao Zhang, Fei Li, Donghong Ji
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add MMCM (missing-modality compensation) (ablation) | Ours · M3D English avg | F1: 53.03 -> 53.80 | Improved |
| add MMCM (ablation) | Ours · M3D Chinese avg | F1: 53.11 -> 53.77 | Improved |
| add DFFM (hierarchical modality fusion) (ablation) | Ours · M3D English avg | F1: 52.78 -> 53.80 | Improved |
| add DFFM (ablation) | Ours · M3D Chinese avg | F1: 51.86 -> 53.77 | Improved |
| add DFFM (ablation) | Ours · M3D Chinese relation extraction | F1: 31.80 -> 36.89 | Improved |
| add MMCM (ablation) | Ours · M3D English relation extraction | F1: 28.17 -> 29.78 | Improved |
| add hierarchical fusion over strongest baseline MDocRE-HN | MDocRE-HN · M3D English avg | F1: 52.65 -> 53.80 | Improved |
| add hierarchical fusion over strongest baseline MDocRE-HN | MDocRE-HN · M3D Chinese avg | F1: 52.42 -> 53.77 | Improved |
| use Ours instead of ViViT+Longformer+TS | ViViT+Longformer+TS · M3D English entity recognition | F1: 92.03 -> 91.24 | Hurt |
| use zero-shot Video-ChatGPT instead of trained model | Video-ChatGPT · M3D English avg | F1: 53.80 -> 22.47 | Hurt |
| use two-stream instead of one-stream fusion | ViViT+Longformer · M3D English avg | F1: 50.99 -> 52.66 | Improved |
| use ViViT instead of VideoMAE with OS fusion | Longformer+OS · M3D Chinese avg | F1: 47.78 -> 50.92 | Improved |
Closest papers by meaning
- A Hierarchical Network for Multimodal Document-Level Relation Extraction (AAAI 2024 · similarity 0.88)
- A Dual-Way Enhanced Framework from Text Matching Point of View for Multimodal Entity Linking (AAAI 2024 · similarity 0.86)
- UMIE: Unified Multimodal Information Extraction with Instruction Tuning (AAAI 2024 · similarity 0.86)
- Prototype-Guided Multimodal Relation Extraction based on Entity Attributes (AAAI 2025 · similarity 0.86)
- Docopilot: Improving Multimodal Models for Document-Level Understanding (CVPR 2025 · similarity 0.85)
- CoMM: A Coherent Interleaved Image-Text Dataset for Multimodal Understanding and Generation (CVPR 2025 · similarity 0.85)
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.
