Research map / Large language models / Retrieval and knowledge
Iterative Multi-Granular RAG with Contextual Hierarchical Graph
Yanli Hu, Teng Liu, Zhuangyi Zhou, Weixin Zeng, Zhen Tan, Xiang Zhao
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 Contextual Hierarchical Graph (ablation) | Naive RAG · 2Wiki and HotpotQA | Recall@5: 2Wiki 61.5 -> 80.1; HotpotQA 75.3 -> 97.0 | Improved |
| add Contextual Hierarchical Graph (ablation) | Naive RAG · PopQA | Recall@5: 56.6 -> 56.6 | No clear effect |
| add query decomposition (ablation) | Naive RAG + CHG · PopQA, 2Wiki, HotpotQA | Recall@5: PopQA 56.6 -> 58.7; 2Wiki 80.1 -> 84.5; HotpotQA 97.0 -> 97.4 | Improved |
| add evidence-augmented reranking (ablation) | MGranRAG · PopQA, 2Wiki, HotpotQA | Recall@5: PopQA 58.7 -> 72.9; 2Wiki 84.5 -> 93.7; HotpotQA 97.4 -> 98.2 | Improved |
| increase iterations from T=1 to T=3 (ablation) | MGranRAG (Qwen3-8B) · average of 6 benchmarks | Recall@5: 83.3 -> 84.2 | Improved |
| increase candidate passage pool M (ablation) | MGranRAG · 2Wiki (T=1) | Recall@5: around 96% at M=10/15 | Improved |
| use NV-Embed-v2 instead of Contriever retriever (ablation) | MGranRAG · 2Wiki | Recall@5 | Improved |
| use MGranRAG instead of HippoRAG 2 for QA | MGranRAG · average F1 6 benchmarks | F1: 55.0 -> 56.2 | Improved |
| increase iterations T=1 to T=3 (ablation) | MGranRAG · LLM token cost on 2Wiki | retrieval input tokens: 4.5M -> 12.0M | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- NV-Embed-v2 — NV-Embed: Improved Techniques for Training LLMs as Generalist Embedding Models (ICLR 2025) · better on 7 of 7 shared results
- HippoRAG 2 — HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Models (NeurIPS 2024) · better on 6 of 7 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
- Simple is Effective: The Roles of Graphs and Large Language Models in Knowledge-Graph-Based Retrieval-Augmented Generation (ICLR 2025 · similarity 0.91)
- Think-on-Graph 2.0: Deep and Faithful Large Language Model Reasoning with Knowledge-guided Retrieval Augmented Generation (ICLR 2025 · similarity 0.91)
- GFM-RAG: Graph Foundation Model for Retrieval Augmented Generation (NeurIPS 2025 · similarity 0.90)
- Can Knowledge-Graph-based Retrieval Augmented Generation Really Retrieve What You Need? (NeurIPS 2025 · similarity 0.90)
- A Systematic Exploration of Knowledge Graph Alignment with Large Language Models in Retrieval Augmented Generation (AAAI 2025 · similarity 0.90)
- HyperGraphRAG: Retrieval-Augmented Generation via Hypergraph-Structured Knowledge Representation (NeurIPS 2025 · similarity 0.89)
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.
