Fast Think-on-Graph: Wider, Deeper and Faster Reasoning of Large Language Model on Knowledge Graph

2025-01-24

AI Summary

Existing Graph-based Retrieval-Augmented Generation (GRAG) methods face three key bottlenecks: shallow and narrow relational modeling, low reasoning efficiency, and weak interpretability—especially on complex queries and large-scale knowledge graphs. To address these, we propose a community-level stepwise reasoning paradigm and introduce the first two-stage graph reasoning framework operating at the community granularity. Stage I performs global topological awareness via community detection and coarse-grained pruning; Stage II executes fine-grained pruning and structure-aware Community-to-Text encoding to realize dual-path subgraph-to-text transformation, tightly coupled with large language model (LLM) co-reasoning. Evaluated on multi-task benchmarks, our method achieves significant gains in both accuracy and inference speed—reducing graph query latency by over 30%—while generating answers grounded in explicit, human-readable subgraph evidence. The approach thus delivers strong interpretability, high scalability, and robust performance across diverse GRAG scenarios.

Technology Category

• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Knowledge Representation and Reasoning: Computational Complexity of Reasoning
• Machine Learning: Graph-based Machine Learning

Application Category

• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Graph Retrieval Augmented Generation (GRAG) is a novel paradigm that takes the naive RAG system a step further by integrating graph information, such as knowledge graph (KGs), into large-scale language models (LLMs) to mitigate hallucination. However, existing GRAG still encounter limitations: 1) simple paradigms usually fail with the complex problems due to the narrow and shallow correlations capture from KGs 2) methods of strong coupling with KGs tend to be high computation cost and time consuming if the graph is dense. In this paper, we propose the Fast Think-on-Graph (FastToG), an innovative paradigm for enabling LLMs to think ``community by community"within KGs. To do this, FastToG employs community detection for deeper correlation capture and two stages community pruning - coarse and fine pruning for faster retrieval. Furthermore, we also develop two Community-to-Text methods to convert the graph structure of communities into textual form for better understanding by LLMs. Experimental results demonstrate the effectiveness of FastToG, showcasing higher accuracy, faster reasoning, and better explainability compared to the previous works.

Problem

• Knowledge Graph Retrieval
• Complex Problem Solving
• Community Interrelation Analysis

Innovation

• FastToG
• Knowledge Graph Retrieval
• Textual Information Transformation