Tag: graph neural networks
All the talks with the tag "graph neural networks".
The Expressive Power of GNNs
Pankaj KumarPublished: at 02:00 PMThis talk delves into the theoretical foundations of Graph Neural Networks (GNNs), focusing on their ability to model complex relationships between node features and graph structures. While GNNs excel in tasks like node classification and link prediction, their expressive power remains challenging to quantify due to the interplay of graph topology and node features. The discussion highlights key challenges, such as trade-offs between expressiveness and computational efficiency, and explores advancements like the connection between GNNs and the Weisfeiler-Lehman graph isomorphism test. The talk also outlines future directions for developing rigorous evaluation methods to better understand and assess GNN expressivity.
Graph Neural Networks - Privacy and Applications
Rucha Bhalchandra JoshiPublished: at 02:00 PMThis work discusses the complex relationships within graph-structured data and the use of graph neural networks (GNNs) for tasks such as node classification and link prediction. It also addresses privacy concerns in GNNs and presents a privacy-preserving approach that safeguards local graph structures while enabling meaningful analysis and insights.
Characterizing Graph Datasets for Node Classification - Homophily-Heterophily Dichotomy and Beyond
Sikta MohantyPublished: at 02:00 PMThis work explores the concept of homophily in graph datasets and proposes a measure called adjusted homophily. It also introduces a new characteristic called label informativeness (LI) to distinguish different types of heterophily. The study shows that LI better correlates with graph neural network performance compared to traditional homophily measures.