On the Evolution of Knowledge Graphs: A Survey and Perspective

TL;DR

This paper provides a comprehensive evolution-driven survey of knowledge graphs (KGs), tracing a progression from Static KG (SKG) to Dynamic KG (DKG), Temporal KG (TKG), and Event KG (EKG). It surveys knowledge extraction and reasoning techniques tailored to each KG type, highlighting static, dynamic, temporal, and event-centric methods, and culminates with practical applications, notably a finance case study that demonstrates KG construction, reasoning, and integration with large language models (LLMs) for financial QA. The authors offer forward-looking perspectives on knowledge representation, extraction, and reasoning, emphasizing neural-symbolic hybrids and the potential integration of KGs with LLMs to enhance robustness, interpretability, and real-time decision making. The work outlines concrete directions such as zero-shot/unsupervised extraction, quality maintenance, non-static reasoning, and EKGR, all aimed at enabling scalable, adaptive, and trustworthy knowledge engineering. Overall, the paper provides actionable taxonomy, methodology comparisons, and strategic insights for advancing KG-driven AI in real-world settings, with a clear roadmap for combining structured KG data with the generative and contextual capabilities of LLMs.

Abstract

Knowledge graphs (KGs) are structured representations of diversified knowledge. They are widely used in various intelligent applications. In this article, we provide a comprehensive survey on the evolution of various types of knowledge graphs (i.e., static KGs, dynamic KGs, temporal KGs, and event KGs) and techniques for knowledge extraction and reasoning. Furthermore, we introduce the practical applications of different types of KGs, including a case study in financial analysis. Finally, we propose our perspective on the future directions of knowledge engineering, including the potential of combining the power of knowledge graphs and large language models (LLMs), and the evolution of knowledge extraction, reasoning, and representation.

Figures (15)

• Figure 1: Examples of four categories of the knowledge graph, i.e., static, dynamic, temporal, and event knowledge graph.
• Figure 2: General framework of the survey.
• Figure 3: The formal evolution history of KGs.
• Figure 4: Examples of a static knowledge graph.
• Figure 5: Examples of a dynamic knowledge graph.