Figure 2 - uploaded by Yaxin Cui
Content may be subject to copyright.
Download
View publication
Illustration of sampling and aggregation in GraphSAGE method. A sample of neighboring nodes contributes to the embedding of the central node.

Illustration of sampling and aggregation in GraphSAGE method. A sample of neighboring nodes contributes to the embedding of the central node.

Source publication
Figure 2: Illustration of sampling and aggregation in GraphSAGE method....
Figure 3: AUC-ROC curve to predict 2014 coconsideration network with 6...
Comparison of this work with prior studies on modeling car relationship...
Confusion matrix in predicting 2014 with six features and 296 cars for...
A Graph Neural Network Approach for Product Relationship Prediction
Preprint
Full-text available
  • May 2021
Graph Neural Networks have revolutionized many machine learning tasks in recent years, ranging from drug discovery, recommendation systems, image classification, social network analysis to natural language understanding. This paper shows their efficacy in modeling relationships between products and making predictions for unseen product networks. By...
Cite
Download full-text

Contexts in source publication

Context 1
... GraphSAGE is a representation learning technique for dynamic graphs. It can predict the embedding of a new node, without needing a re-training procedure. To do this, GraphSAGE uses inductive learning. It learns aggregator functions that can induce new node embedding, based on the features and neighborhood of the node. As illustrated in Fig. 2, GraphSAGE learns node embeddings for attributed graphs (where nodes have features or attributes) through aggregating neighboring node attributes. The aggregation parameters are learned by encouraging node pairs co-occurring in short random walks to have similar representations. Many GNN models learn functions that generate the ...
Context 2
... GraphSAGE is a representation learning technique for dynamic graphs. It can predict the embedding of a new node, without needing a re-training procedure. To do this, GraphSAGE uses inductive learning. It learns aggregator functions that can induce new node embedding, based on the features and neighborhood of the node. As illustrated in Fig. 2, GraphSAGE learns node embeddings for attributed graphs (where nodes have features or attributes) through aggregating neighboring node attributes. The aggregation parameters are learned by encouraging node pairs co-occurring in short random walks to have similar representations. Many GNN models learn functions that generate the ...