Knowledge graph completion method and device, electronic equipment and storage medium

• Main Authors: XU HUIFANG, XU JIANNAN, HE SHIZHU, JIAO FEI, ZHAO JUN, ZHANG CHUNHUI, LI SONGYUAN, LIU LIQING, LIU KANG, LI LONGJI, TANG QINGHUA, ZHENG QU'AN, LAN YINYU
• Format: Patent
• Language: Chinese; English
• Published: 26.05.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention provides a knowledge graph completion method and device, electronic equipment and a storage medium, and relates to the technical field of natural language processing, and the method comprises the steps: obtaining a plurality of fact triples in a knowledge graph in a current iteration period; determining at least one conclusion triple based on logic rules among the fact triads; the candidate triads and labels corresponding to the candidate triads are input into a joint representation learning model, vector representation and a final score value corresponding to the candidate triads are obtained, and a loss function of the joint representation learning model is determined based on conclusion loss of logic rules, minimization of global loss and L2 norm; and determining a reasoning triple based on the conclusion triple, updating the knowledge graph based on the reasoning triple, and performing iterative representation learning based on the updated knowledge graph until the conclusion triple cannot b