Research map / Large language models / Retrieval and knowledge
Geometric Gain Graph: Zero-Token Graph Construction for Multi-Hop RAG
Zeliang Li, Xiaofen Xing, Wenyu Tao, Kailing Guo, Xiangmin Xu
Read the paperCheck your idea against 137,217 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes Geometric Gain Graph, a zero-token graph construction method for multi-hop RAG that balances relevance and novelty without LLM-based entity extraction.
Problem
Multi-hop reasoning in RAG fails because dense retrievers collapse onto similar surface-level documents (the "Similarity Trap"), while Graph RAG methods explore novel documents via heuristic entity extraction at prohibitive LLM cost.
Approach
Builds the retrieval graph geometrically from embeddings so that relevance and novelty are balanced without any LLM calls for entity extraction, giving zero-token graph construction.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add topological degree penalty to graph edges (ablation) | G3RAG · MuSiQue (Nv-embed-v2) | F1: 50.65 -> 52.28 | Improved |
| add topological degree penalty to graph edges (ablation) | G3RAG · HotpotQA (Nv-embed-v2) | F1: 77.60 -> 78.00 | Improved |
| use single-step controlled diffusion instead of Personalized PageRank (ablation) | G3RAG · Average of 3 datasets (Nv-embed-v2) | F1: 65.65 -> 66.49 | Improved |
| add LLM filtering of seed nodes (ablation) | G3RAG · Average of 3 datasets (Nv-embed-v2) | F1: 60.53 -> 66.49 | Improved |
| add geometric graph diffusion on top of LLM-filtered naive retrieval (ablation) | Naive RAG with Filter_llm · Average of 3 datasets (Nv-embed-v2) | F1: 63.39 -> 66.49 | Improved |
| use single-step diffusion instead of greedy walk or beam search (ablation) | G3RAG · Average of 3 datasets (Nv-embed-v2) | F1: 64.79 (greedy), 64.87 (beam) -> 66.49 | Improved |
| use global density estimate instead of local 10-NN density for penalty (ablation) | G3RAG · Average of 3 datasets (Nv-embed-v2) | F1: 64.97 -> 66.49 | Improved |
| use geometric gain graph instead of MMR or DPP diversification | Naive RAG · Average of 3 datasets (Nv-embed-v2) | F1: MMR 52.47, DPP 60.93 -> 66.49 | Improved |
| add MMR diversification to naive dense retrieval | Naive RAG · Average of 3 datasets (Nv-embed-v2) | F1: 60.62 -> 52.47 | Hurt |
| add DPP diversification to naive dense retrieval | Naive RAG · Average of 3 datasets (Nv-embed-v2) | F1: 60.62 -> 60.93 | No clear effect |
| plug G3RAG in as retriever of iterative frameworks | IterDRAG and IterRetgen · Average of 3 datasets (Nv-embed-v2) | F1: IterDRAG 62.54 -> 66.26; IterRetgen 64.25 -> 68.42 | Improved |
| reduce HippoRAG2 seed count from 5 to 2 (ablation) | HippoRAG2 · Average of 3 datasets (Nv-embed-v2) | F1: 64.49 -> 63.21 | 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.
- HippoRAG2 — HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Models (NeurIPS 2024) · better on 6 of 8 shared results
- LinearRAG — LinearRAG: Linear Graph Retrieval Augmented Generation on Large-scale Corpora (ICLR 2026) · better on 1 of 1 shared result
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.
Experimental setup
Baselines: dense retrievers Graph RAG
Builds on: Graph RAG
Closest papers by meaning
- GFM-RAG: Graph Foundation Model for Retrieval Augmented Generation (NeurIPS 2025 · similarity 0.89)
- Can Knowledge-Graph-based Retrieval Augmented Generation Really Retrieve What You Need? (NeurIPS 2025 · similarity 0.88)
- A Systematic Exploration of Knowledge Graph Alignment with Large Language Models in Retrieval Augmented Generation (AAAI 2025 · similarity 0.88)
- HyperGraphRAG: Retrieval-Augmented Generation via Hypergraph-Structured Knowledge Representation (NeurIPS 2025 · similarity 0.87)
- RAGraph: A General Retrieval-Augmented Graph Learning Framework (NeurIPS 2024 · similarity 0.87)
- Simple is Effective: The Roles of Graphs and Large Language Models in Knowledge-Graph-Based Retrieval-Augmented Generation (ICLR 2025 · similarity 0.87)
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.
