Unifying Large Language Models and Knowledge Graphs: A Roadmap

By Shirui Pan, Linhao Luo, Yufei Wang, Chen Chen, Jiapu Wang, Xindong Wu. IEEE Transactions on Knowledge and Data Engineering, 2023

1,606 citations recorded by the source corpus.

Large language models (LLMs), such as ChatGPT and GPT4, are making new waves in the field of natural language processing and artificial intelligence, due to their emergent ability and generalizability. However, LLMs are black-box models, which often fall short of capturing and accessing factual knowledge. In contrast, Knowledge Graphs (KGs), Wikipedia, and Huapu for example, are structured knowledge models that explicitly store rich factual knowledge. KGs can enhance LLMs by providing external knowledge for inference and interpretability. Meanwhile, KGs are difficult to construct and evolve by nature, which challenges the existing methods in KGs to generate new facts and represent unseen knowledge. Therefore, it is complementary to unify LLMs and KGs together and, simultaneously, leverage their advantages. In this article, we present a forward-looking roadmap for the unification of LLMs and KGs. Our roadmap consists of three general frameworks, namely: 1) KG-enhanced LLMs, which incorporate KGs during the pre-training and inference phases of LLMs, or for the purpose of enhancing understanding of the knowledge learned by LLMs; 2) LLM-augmented KGs, that leverage LLMs for different KG tasks such as embedding, completion, construction, graph-to-text generation, and question answering; and 3) Synergized LLMs + KGs, in which LLMs and KGs play equal roles and work in a mutually beneficial way to enhance both LLMs and KGs for bidirectional reasoning driven by both data and knowledge. We review and summarize existing efforts within these three frameworks in our roadmap and pinpoint their future research directions.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/9e8b7b0d4c628c12b6a65ab56ac5f33a35eff2e6

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Shirui Pan
2. Heng Ji
3. Ming Jin
4. Xuanjing Huang
5. Jiawei Han
6. Yue Liu
7. Siwei Wang
8. Lingyuan Meng

Citations in the corpus

• Think-on-Graph: Deep and Responsible Reasoning of Large Language Model with Knowledge Graph (2023, background): “…and catastrophic forgetting problems produced during fine-tuning or continual pre-training of LLM. Knowledge graphs (KGs) provide high-quality and explicit knowledge representations, and are easily editable, thus being considered as a promising complement to the drawbacks of LLMs Pan et al. (2023).”
• A Survey on Evaluation of Large Language Models (2023, neutral): “These include tasks such as mathematical reasoning [214, 224, 230] and structured data inference [79, 141].”
• Large Language Models on Graphs: A Comprehensive Survey (2023, methodology): “Pan et al. [21] review the connection between LLMs and knowledge graphs (KGs) especially on how KGs can enhance LLMs training and inference, and how LLMs can facilitate KG construction and reasoning.”
• Reasoning on Graphs: Faithful and Interpretable Large Language Model Reasoning (2023, background): “Knowledge graphs contain abundant factual knowledge in a structured format, which has attracted great attention from researchers (Zheng et al., 2022a; 2023a; Pan et al., 2023; Liang et al., 2023).”
• Think-on-Graph: Deep and Responsible Reasoning of Large Language Model on Knowledge Graph (2023, neutral): “KGs offer structured, explicit, and editable representations of knowledge, presenting a complementary strategy to mitigate the limitations of LLMs (Pan et al., 2023).”
• The rise and potential of large language model based agents: a survey (2025, positive): “For example, search-based tools can improve the scope and quality of the knowledge accessible to the agents with the aid of external databases, knowledge graphs, and web pages, while domain-specific tools can enhance an agent’s expertise in the corresponding field [243,244].”
• Software Testing With Large Language Models: Survey, Landscape, and Vision (2023, neutral): “In another survey of unifying LLMs and knowledge graphs [24], the authors categorize the LLMs into three types: encoder-only (e.g., BERT), encoder-decoder (e.g., T5), and decoder-only network architecture (e.g., GPT-3).”
• A Survey on Graph Neural Networks for Time Series: Forecasting, Classification, Imputation, and Anomaly Detection (2023, background): “These techniques hinge on utilizing learned representations from one or more domains to enhance performance in other related domains [3], [289].”
• When large language models meet personalization: perspectives of challenges and opportunities (2023, neutral): “For example, APIs for querying knowledge graphs [236] or accessing users’ social relationships can”
• A Survey of Knowledge Graph Reasoning on Graph Types: Static, Dynamic, and Multi-Modal (2022, background): “, ChatGPT, GPT-4, are very popular this year, which have huge impacts due to their promising reasoning capacity and generalizability [263].”
• Large Language Models for Software Engineering: A Systematic Literature Review (2023, methodology): “In this paper, we adopt the LLM scope division and taxonomy introduced by Pan et al.[216] and categorize the mainstream LLMs investigated in this study into three groups according to their architectures: encoder-only, encoder-decoder, and decoder-only LLMs.”
• MARS: Multi-hop Adaptive Retrieval and SPARQL Generation for KGQA (2026, unknown): “Recent work explores the integration of KGs in LLMs for KGQA [23,38,49].”
• GraphRAG on Consumer Hardware: Benchmarking Local LLMs for Healthcare EHR Schema Retrieval (2026, unknown): “Pan et al. [19] outline multiple paradigms for combining symbolic and neural representations, including LLM-augmented knowledge graphs where models are used to construct structured representations from text.”
• An Empirical Study of Knowledge Graph-Enhanced RAG for Information Security Compliance (2026, unknown): “Integration of knowledge graphs with large language models has emerged as a promising research direction, with recent work demonstrating that explicit structural representations can improve reasoning, reduce hallucination, and enable more interpretable output [27,28].”
• Higher-Order Knowledge Representations for Agentic Scientific Reasoning (2026, unknown): “Recent work has shown that augmenting LLMs with graph-based and ontological knowledge representations facilitates compositional reasoning across disparate knowledge domains and the discovery of previously unseen relationships [14, 26, 40, 41].”
• Retrieval And Structuring Augmented Generation with Large Language Models (2025, unknown): “These approaches enable LLMs to leverage the structured representations in KGs more effectively, improving reasoning capabilities while maintaining interpretability.”