Year: 2024
Type: dissertation
Abstract: In this thesis, the author aims to develop advanced graph neural networks (GNNs) to learn entity representations within real-world graph-structured datasets.The developed GNNs aim to address predictive analytics on tasks in human behaviors and social trends activities.This thesis mainly focuses on three research questions: How to address prediction tasks on heterogeneous graphs?How to address prediction tasks on dynamic graphs?Finally, how to tailor GNNs for domain-specific prediction tasks.The author conducted three subworks to address the research questions and discussed the relationships within them.
Authors: JIN, Ruidong, Jin, Ruidong
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6995897848
