Impact Statement:Modeling complex data correlations is important to understand hidden and rich semantics in real-world data. Graph learning has been powerful for capturing correlation in ...Show More
Abstract:
Graphs are a foundational way to represent scenarios where objects interact in pairs. Recently, graph neural networks (GNNs) have become widely used for modeling simple g...Show MoreMetadata
Impact Statement:
Modeling complex data correlations is important to understand hidden and rich semantics in real-world data. Graph learning has been powerful for capturing correlation in pairwise data and is widely applied for tasks such as node classification and link prediction. However, real-world data often involves interactions that go beyond simple pairwise relationships, which traditional graphs cannot adequately represent. Hypergraph learning addresses this limitation by modeling complex and high-order data correlations. While recent research on hypergraph learning has predominantly focused on static and homogeneous data, it is important to recognize that real-world data inherently exhibits two significant properties – heterogeneous and dynamic. The method proposed in this article is tailored to model high-order complex correlations for the node classification task in dynamic networks. Our method achieves a 2.75% improvement in overall performance compared with conventional methods, making it w...
Abstract:
Graphs are a foundational way to represent scenarios where objects interact in pairs. Recently, graph neural networks (GNNs) have become widely used for modeling simple graph structures, either in homogeneous or heterogeneous graphs, where edges represent pairwise relationships between nodes. However, many real-world situations involve more complex interactions where multiple nodes interact simultaneously, as observed in contexts such as social groups and gene-gene interactions. Traditional graph embeddings often fail to capture these multifaceted nonpairwise dynamics. A hypergraph, which generalizes a simple graph by connecting two or more nodes via a single hyperedge, offers a more efficient way to represent these interactions. While most existing research focuses on homogeneous and static hypergraph embeddings, many real-world networks are inherently heterogeneous and dynamic. To address this gap, we propose a GNN-based embedding for dynamic heterogeneous hypergraphs, specifically d...
Published in: IEEE Transactions on Artificial Intelligence ( Volume: 5, Issue: 11, November 2024)