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.