Abstract
Graph Neural Networks (GNNs) have been widely used in various graph-related tasks and exhibited competitive performance. Existing GNNs follow a message-passing rule that aggregates the information of neighbors to update node representations. The design of message-passing function is the most fundamental part of GNNs. In this chapter, we will introduce the message-passing functions of three representative homogeneous GNNs. Further, we show that most existing homogeneous GNNs can be unified as a closed-form framework, which may help the researchers understand and interpret the principles behind message-passing mechanism.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this chapter
Cite this chapter
Bo, D. (2023). Homogeneous Graph Neural Networks. In: Advances in Graph Neural Networks. Synthesis Lectures on Data Mining and Knowledge Discovery. Springer, Cham. https://doi.org/10.1007/978-3-031-16174-2_3
Download citation
DOI: https://doi.org/10.1007/978-3-031-16174-2_3
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-16173-5
Online ISBN: 978-3-031-16174-2
eBook Packages: Synthesis Collection of Technology (R0)eBColl Synthesis Collection 12