In recent years, Knowledge Graphs (KGs) have played a crucial role in the development of advanced knowledge-intensive applications, such as recommender systems and semantic search. However, the human sensory system is inherently multi-modal, as objects around us are often represented by a combination of multiple signals, such as visual and textual. Consequently, Multi-modal Knowledge Graphs (MMKGs), which combine structured knowledge representation with multiple modalities, represent a powerful extension of KGs. Although MMKGs can handle certain types of tasks (e.g., visual query answering) or queries that standard KGs cannot process, and they can effectively tackle some standard problems (e.g., entity alignment), we lack a widely accepted definition of MMKG. In this survey, we provide a rigorous definition of MMKGs along with a classification scheme based on how existing approaches address four fundamental challenges: representation, fusion, alignment, and translation, which are crucial to improving an MMKG. Our classification scheme is flexible and allows for easy incorporation of new approaches, as well as a comparison of two approaches in terms of how they address one of the fundamental challenges mentioned above. As the first comprehensive survey of MMKG, this article aims at inspiring and provide a reference for relevant researchers in the field of Artificial Intelligence.

Action recognition based on 3D skeleton data has attracted much attention due to its wide application, and it is one of the most popular research topics in computer vision. The 3D skeleton data is an effective representation of motion dynamics and is not easily affected by light, scene variation, etc. Previous research on action recognition has mainly focused on video or RGB data methods. In recent years, the advantages of combining skeleton data and deep learning have been gradually demonstrated, many impressive methods have been proposed, especially GCN‐based methods. In this survey, we first introduce the development process of 3D skeleton‐data action recognition and the classification of graph convolutional network, then introduce the commonly used NTU RGB + D and NTU RGB + D 120 datasets. Finally, a detailed review of existing variants of three mainstream technologies is provided based on deep learning and their performance was compared from three dimensions. To the best of our knowledge, this is the first research to integrate the research of GCN‐based method and its various evolutionary methods. Comparative investigation of existing variants of research in action‐recognition task from different perspectives is made, a generic framework is described, state‐of‐the‐art practices are summarized, and the emerging trends of this topic are explored.