Article Preview
TopIntroduction
At present, the rapid development of many fields cannot be separated from power. However, with the innovation and development of energy management systems worldwide, the traditional power field also needs to transform to informationized power; thus, the innovation and transformation of power systems is imperative. Knowledge Graphs (KGs) are an intelligent database system that combines artificial intelligence with traditional databases. It can realise structural management of large-scale knowledge. By integrating KGs with the power professional field, the scattered knowledge points in the power field can be connected. This study proposes an automatic completion algorithm of power professional KGs on the basis of convolutional neural networks (CNNs). This study hopes to provide support for electric power companies to grasp the development trend and construction of the field firmly.
Research on KGs has always attracted the attention of many experts and scholars. Such research plays an important role in intelligent search (It refers to the one-stop search for all mainstream resources that can be found on the Internet such as web pages, music, games, pictures, movies, and shopping) and decision-making. At present, Hogan (2021) has made a comprehensive introduction to KGs, compared various data models in view of the graphs, and compared the languages used to query and verify KGs. He explained how to use a combination of deductive and inductive techniques to represent and extract knowledge. Finally, he gave the advanced research direction of Knowledge Graphing for the future (Hogan, 2021). Ji (2021) comprehensively reviewed KGs, covering the overall research topics, including KG representation learning, KG knowledge acquisition and completion, time KGs, and knowledge perception application. He reviewed embedding methods, path reasoning, and logic rule reasoning. To promote future research on KGs, he also provided a planning set of datasets and open-source libraries for different tasks (Ji, 2021). Wang (2017) introduced the technology of embedding using only the facts observed in the KGs. He described the overall framework, specific model design, typical training procedures, and advantages and disadvantages of these technologies. He briefly introduced how KG embedding was applied to various downstream tasks, such as KG completion, relationship extraction, question answering, and so on (Wang, 2017). Noy (2019) focused on the KG of five different technology companies and compared their similarities and differences in building and using these maps. He discussed the challenges facing all knowledge-driven enterprises. The KG set discussed here consists of the products searched and the social networks described (Noy, 2019). These studies play an important role in promoting the development of KGs.