METHOD AND DEVICE FOR TRAINING GRAPH NEURAL NETWORK MODEL

摘要

Provided by embodiments of the present description are a method and device for training a graph neural network model. The method comprises: acquiring a target training sample and a corresponding target sample tag from within a sample set, wherein the target training sample corresponds to a target node in a target relational network graph, the target node has a target node number, the target relational network graph comprises multiple nodes and connection sides between the nodes, each node has a respective corresponding node number, and each connection side has a respective corresponding side number; searching for graph information of a target subgraph of the target relational network graph from among graph information of a prestored target relational network graph according to the target node number and a preset parameter, wherein the target subgraph uses the target node as a central node, and the hop count between each node in the target subgraph and the target node is less than or equal to the preset parameter; training a graph neural network model by using the graph information of the target subgraph and the target sample tag. Requirements on a machine may be reduced, and the training efficiency is improved.