Abstract
Graph Neural Networks (GNNs) excel at learning from graph-structured data butare limited to modeling pairwise interactions, insufficient for capturinghigher-order relationships present in many real-world systems. Topological DeepLearning (TDL) has allowed for systematic modeling of hierarchical higher-orderinteractions by relying on combinatorial topological spaces such as simplicialcomplexes. In parallel, Quantum Neural Networks (QNNs) have been introduced toleverage quantum mechanics for enhanced computational and learning power. Inthis work, we present the first Quantum Topological Deep Learning Model:Quantum Simplicial Networks (QSNs), being QNNs operating on simplicialcomplexes. QSNs are a stack of Quantum Simplicial Layers, which are inspired bythe Ising model to encode higher-order structures into quantum states.Experiments on synthetic classification tasks show that QSNs can outperformclassical simplicial TDL models in accuracy and efficiency, demonstrating thepotential of combining quantum computing with TDL for processing data oncombinatorial topological spaces.