Knowledge graph fusion method based on integrated semantic rule and representation learning

• Main Authors: CHU TAO, ZHU YUHU, WANG CHAOJUN, ZHANG JINGANG, ZHANG XINGXING, SUN LIQIANG, CHENG PUQIANG, FENG LI, HAN BING, WANG QIBIN
• Format: Patent
• Language: Chinese; English
• Published: 08.03.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention discloses a knowledge graph fusion method based on an integrated semantic rule and representation learning, and mainly solves the problems that in the prior art, equivalent triads cannot be effectively fused under the condition of few sample data, and the fusion generalization is low. According to the implementation scheme, firstly, a knowledge fusion model based on representation learning is trained, then primary knowledge fusion is carried out through a semantic rule, and a fusion counter example and a fusion positive example after primary fusion are obtained; inputting the fusion counter example into a representation-based learning knowledge fusion model for secondary fusion to obtain a fusion counter example and a fusion positive example after secondary fusion; and finally, integrating positive examples in the two fusion results to complete the fusion alignment of the equivalent triad. According to the method, the accuracy of equivalent triple fusion can be improved under the condition of fe