Research map / Large language models / Retrieval and knowledge
LeanRAG: Knowledge-Graph-Based Generation with Semantic Aggregation and Hierarchical Retrieval
Yaoze Zhang, Rong Wu, Pinlong Cai, Xiaoman Wang, Guohang Yan, Song Mao, Ding Wang, Botian Shi
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 original text chunks to the generator context (ablation) | LeanRAG · UltraDomain Mix | Overall score: 7.93 -> 8.59 | Improved |
| add original text chunks to the generator context (ablation) | LeanRAG · UltraDomain CS | Overall score: 8.34 -> 8.82 | Improved |
| add original text chunks to the generator context (ablation) | LeanRAG · UltraDomain Agriculture | Overall score: 8.53 -> 8.87 | Improved |
| add explicit inter-cluster aggregated relations (ablation) | LeanRAG · UltraDomain CS, Diversity win rate | win rate: 66.0% vs 34.0% | Improved |
| add explicit inter-cluster aggregated relations (ablation) | LeanRAG · UltraDomain Mix, Comprehensiveness win rate | win rate: 51.5% vs 48.6% | No clear effect |
| add explicit inter-cluster aggregated relations (ablation) | LeanRAG · UltraDomain Mix, Overall win rate | win rate: 53.8% vs 46.2% | Improved |
| LCA-based bottom-up hierarchical retrieval instead of community retrieval | LeanRAG · Four UltraDomain datasets, Qwen3-14B | retrieved context tokens: 46% smaller on average | Improved |
| switch from HiRAG to LeanRAG | KG-based RAG · UltraDomain Mix | Overall score: 8.08 -> 8.59 | Improved |
| switch from NaiveRAG/HiRAG to LeanRAG | KG-based RAG · UltraDomain CS and Agriculture | Overall score: CS 8.82 vs 8.77; Agriculture 8.87 vs 8.87 | No clear effect |
| switch from GraphRAG to LeanRAG | KG-based RAG · UltraDomain Legal, Comprehensiveness | Comprehensiveness score: 8.95 -> 8.88 | Hurt |
Closest papers by meaning
- Think-on-Graph 2.0: Deep and Faithful Large Language Model Reasoning with Knowledge-guided Retrieval Augmented Generation (ICLR 2025 · similarity 0.92)
- Simple is Effective: The Roles of Graphs and Large Language Models in Knowledge-Graph-Based Retrieval-Augmented Generation (ICLR 2025 · similarity 0.91)
- HyperGraphRAG: Retrieval-Augmented Generation via Hypergraph-Structured Knowledge Representation (NeurIPS 2025 · similarity 0.91)
- A Systematic Exploration of Knowledge Graph Alignment with Large Language Models in Retrieval Augmented Generation (AAAI 2025 · similarity 0.90)
- SiReRAG: Indexing Similar and Related Information for Multihop Reasoning (ICLR 2025 · similarity 0.90)
- Can Knowledge-Graph-based Retrieval Augmented Generation Really Retrieve What You Need? (NeurIPS 2025 · similarity 0.90)
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.
