Influential Simplices Mining via Simplicial Convolutional Network

Abstract

Simplicial complexes play a critical role in higher-order network analysis due to their heterogeneity, but identifying influential simplices remains elusive. This paper proposes the influential simplices mining neural network (ISMnet), which uses higher-order presentations and graph convolutional operators to identify vital simplices of arbitrary order. Empirical results demonstrate that ISMnet significantly outperforms existing methods in ranking nodes and 2-simplices, making it a potent tool in higher-order network analysis.

Yiming Huang

PhD Candidate

My research focuses on topological deep learning for complex and higher-order graph structures, as well as generative models, graph representation learning, and foundation models.

Linyuan Lü

Professor

Professor, doctoral supervisor, winner of the National Natural Science Foundation of Outstanding Youth Science Fund.