サインイン
インサイト - Simplicial neural networks
Efficient Learning from Simplicial Data Using Random Walks and 1D Convolutions
A novel simplicial neural network architecture, SCRaWl, that leverages random walks and 1D convolutions to efficiently capture higher-order relationships in simplicial data.
Generalized Simplicial Attention Neural Networks: A Novel Architecture for Processing Data on Simplicial Complexes
This paper introduces GSAN, a novel neural network architecture specifically designed to process data structured on simplicial complexes, leveraging the power of masked self-attention mechanisms and principles of topological signal processing for enhanced learning and representation capabilities.
