Injecting Knowledge Graphs into Large Language Models

2025-05-12

AI Summary

To address the weak symbolic reasoning capability of large language models (LLMs) stemming from insufficient structured knowledge support, this paper proposes KG-Inject—a lightweight integration method that directly injects knowledge graph embeddings (KGEs) as learnable tokens into the LLM’s input layer. KG-Inject achieves, for the first time, deep fusion between KGEs and LLM input encoding without modifying model parameters or requiring fine-tuning, thereby preserving knowledge structure fidelity while maintaining computational efficiency. Its model-agnostic design enables plug-and-play compatibility with arbitrary open- or closed-source LLMs. Extensive experiments on both synthetic and real-world datasets demonstrate that KG-Inject significantly improves logical reasoning accuracy, outperforming state-of-the-art knowledge-augmented methods in both overall precision and inference efficiency.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Linked Open Data, Knowledge Graphs & KB Completion
• Knowledge Representation and Reasoning: Knowledge Representation Languages

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

Integrating structured knowledge from Knowledge Graphs (KGs) into Large Language Models (LLMs) remains a key challenge for symbolic reasoning. Existing methods mainly rely on prompt engineering or fine-tuning, which lose structural fidelity or incur high computational costs. Building on recent encoding techniques which integrate graph embeddings within the LLM input as tokens, we extend this paradigm to the KG domain by leveraging Knowledge Graph Embedding (KGE) models, thus enabling graph-aware reasoning. Our approach is model-agnostic, resource-efficient, and compatible with any LLMs. Extensive experimentation on synthetic and real-world datasets shows that our method improves reasoning performance over established baselines, further achieving the best trade-off in terms of accuracy and efficiency against state-of-the-art LLMs.

Problem

• Integrating Knowledge Graphs into LLMs for symbolic reasoning
• Overcoming structural fidelity loss and high computational costs
• Enhancing reasoning performance with graph-aware, resource-efficient methods

Innovation

• Integrates Knowledge Graph Embeddings as LLM tokens
• Model-agnostic and resource-efficient KG-LLM fusion
• Enhances reasoning accuracy with graph-aware inputs