DLGNet: Hyperedge Classification via a Directed Line Graph for Chemical Reactions

ICLR 2026 Conference Submission5148 Authors

Keywords: Directed Line Graph, Directed Line Graph Laplacian, Hyperedge Classification, Chemical reaction classification

Abstract: Graphs and hypergraphs provide powerful abstractions for modeling interactions among a set of entities of interest and have been attracting a growing interest in the literature thanks to many successful applications in several fields, including chemistry. In the paper, we address the reaction classification task by introducing the Directed Line Graph (DLG) transformation for directed hypergraphs. Building on this representation, we propose the Directed Line Graph Network (DLGNet), the first spectral-based Graph Neural Network (GNN) designed to perform convolutions directly on the DLG. At the core of DLGNet lies a novel complex-valued Hermitian matrix, the Directed Line Graph Laplacian ($\mathbb{\vec{L}}_N$), which effectively encodes directional interactions within the hypergraph’s structure through the DLG representation. Experimental results on three real-world chemical reaction datasets demonstrate that DLGNet consistently outperforms all baseline competitors.

Primary Area: applications to physical sciences (physics, chemistry, biology, etc.)

Submission Number: 5148