Knowledge Graph Construction of High-Performance Computing Learning Platform

• Published in: Journal of physics. Conference series Vol. 1748; no. 2; pp. 22035 - 22044
• Main Authors: Dong, Tianyue, Tang, Lei, Peng, Jinye, Zhong, Sheng, Luo, Hangzai
• Format: Journal Article
• Language: English
• Published: Bristol IOP Publishing 01.01.2021
• Subjects: Algorithms; Annotations; Clustering; Computation; Data collection; Feature extraction; Graphs; High performance computing; Knowledge bases (artificial intelligence); Knowledge representation; Learning; Model testing; Modules; Neural networks; Physics; Semantics; Unstructured data
• ISSN: 1742-6588; 1742-6596
• DOI: 10.1088/1742-6596/1748/2/022035

With the development of intelligent education, it has become one of the more efficient learning schemes to construct the knowledge graph which can excavate the knowledge base. People generally use RDF triples and use languages such as OWL to construct knowledge graphs, but this method has problems such as limited expression ability and too much manual annotation. In this paper, we propose a framework that combines statistical language models, neural network language models, and clustering and clipping algorithms. When processing unstructured text data, unsupervised extraction of representative key words as features of the structured graphs' entities, so that the processed entity information has more accurate semantics and human learning relevance, and the method of clustering and calculating the learning rate is used to further clarify the learning order and mastery of knowledge points. We have conducted extensive experiments to collect data from the two major modules of HPC high-performance computing courses and domains as datasets, and used this framework to build many knowledge graphs which can provide practical learning. The knowledge graph of this paper can be obtained from: https://v2.easyhpc.net:10000/knowledge Knowledge Graph Module.