The National Science Foundation (NSF) awarded a $600,000 Project Grant under its Computer and Information Science and Engineering (CFDA 47.070) program to the University of Virginia (UVA) to develop innovative approaches for efficient training of dynamic graph neural networks (DGNNs) on large-scale and time-varying graphs. The 3-year project aims to advance the state-of-the-art in DGNN training by creating novel techniques for graph partitioning, sampling, caching, and training to enable highly scalable and efficient DGNN execution on real-world, dynamic graph datasets. The funded research seeks to address key challenges in leveraging DGNNs for applications in diverse domains, including social networks and natural language processing. The project will also provide training opportunities for graduate, undergraduate, and K-12 students, and integrate the research findings into academic courses.
Generated 3/4/25, 5:42 AM