Research map / Vision-language and multimodal / Multimodal LLMs
MG2-RAG: Multi-Granularity Graph for Multimodal Retrieval-Augmented Generation
Sijun Dai, Qiang Huang, Xiaoxing You, Jun Yu
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 hierarchical graph construction (ablation) | MG2-RAG · E-VQA (5k) | R@1: 40.7 -> 57.8 | Improved |
| add multimodal node fusion (ablation) | MG2-RAG · E-VQA (5k) | BEM: 55.38 -> 60.24 | Improved |
| add graph propagation (ablation) | MG2-RAG · E-VQA (5k) | R@1: 25.5 -> 57.8 | Improved |
| add hierarchical graph construction (ablation) | MG2-RAG · ScienceQA | average accuracy: 97.48 -> 97.85 | No clear effect |
| add hierarchical graph construction (ablation) | MG2-RAG · CrisisMMD | average accuracy: 56.70 -> 59.12 | Improved |
| add multimodal node fusion and graph propagation (ablation) | MG2-RAG · CrisisMMD | average accuracy: 58.62 -> 59.12 | No clear effect |
| add MG2-RAG retrieval to GPT-5.2 | GPT-5.2 · E-VQA | BEM: 40.30 -> 60.30 | Improved |
| add MG2-RAG retrieval to Qwen2.5-VL-7B | Qwen2.5-VL-7B · ScienceQA | average accuracy: 87.50 -> 89.25 | Improved |
| add MG2-RAG retrieval to GPT-5.2 | GPT-5.2 · ScienceQA | average accuracy: 96.30 -> 97.00 | Improved |
| add MG2-RAG retrieval to Qwen2.5-VL-7B | Qwen2.5-VL-7B · CrisisMMD | average accuracy: 51.41 -> 54.91 | Improved |
| use lightweight parsing and grounding instead of MLLM triplet extraction | MG2-RAG · graph construction time | speedup: 43.3x | Improved |
| set propagation alpha to 0.20 instead of 0.50 (ablation) | MG2-RAG · E-VQA (5k) | R@1: 55.5 -> 57.8 | Improved |
Closest papers by meaning
- Navigating Massive Visual Context in Retrieval-Augmented Generation via Multimodal Memory Graph (ICML 2026 · similarity 0.90)
- Beyond Global Similarity: Multi-Conditional Retrieval for Fine-Grained Cross-Modal Understanding (CVPR 2026 · similarity 0.89)
- Simple is Effective: The Roles of Graphs and Large Language Models in Knowledge-Graph-Based Retrieval-Augmented Generation (ICLR 2025 · similarity 0.89)
- A Systematic Exploration of Knowledge Graph Alignment with Large Language Models in Retrieval Augmented Generation (AAAI 2025 · similarity 0.89)
- Think-on-Graph 2.0: Deep and Faithful Large Language Model Reasoning with Knowledge-guided Retrieval Augmented Generation (ICLR 2025 · similarity 0.89)
- ARK: A Dual-Axis Multimodal Retrieval Benchmark along Reasoning and Knowledge (NeurIPS 2026 · similarity 0.88)
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.
