How to Define a Multi-modal Knowledge Graph?

As a form of structured human knowledge, knowledge graphs (KG) have attracted great attention from both the academic and industrial communities since their emergence. It is widely used in the field of artificial intelligence for applications such as information retrieval, data analysis, intelligent question-answering and recommendation systems. In recent years, various types of information on the internet have exploded in growth. In response, multimodal knowledge graphs (MMKGs) have emerged to serve the management and applications of different types of data. However, since the proposal of KG in 2012, there has not been a unified and standardized definition to describe KG, let alone MMKG. Based on previous research and experience, this paper has summarized the definition of KG through extensive investigation and explores the concept of MMKG. To provide a better illustration, this paper constructed a sample MMKG in the medical field based on an ontology and resource description framework (RDF). We use Neo4j for visualization and design a UI to extract node information. Finally, the shortcomings of the work were summarized, and future research directions were proposed.