Bayes EMbedding (BEM): Refining Representation by Integrating Knowledge Graphs and Behavior-specific Networks

• DOI: 10.48550/arxiv.1908.10611
• Publication Date: 2019-01-01
• Authors: Yuting Ye; Xuwu Wang; Jiangchao Yao; Kunyang Jia; Jingren Zhou; Yanghua Xiao; Hongxia Yang
• Keywords: Representation; Benchmark (surveying); Generative model; Link (geometry); Knowledge graph

Low-dimensional embeddings of knowledge graphs and behavior have proved remarkably powerful in varieties tasks, from predicting unobserved edges between entities to content recommendation. The two types can contain distinct complementary information for the same entities/nodes. However, previous works focus either on graph embedding or while few consider both a unified way. Here we present BEM , Bayesian framework that incorporates graphs. To be more specific, takes as prior pre-trained graph, integrates them with via generative model. is able mutually refine sides preserving their own topological structures. show superiority our method, conduct range experiments three benchmark datasets: node classification, link prediction, triplet classification small datasets related Freebase, item recommendation large-scale e-commerce dataset.