Abstract
This overview article makes the case for how topological concepts can enrichresearch in machine learning. Using the Euler Characteristic Transform (ECT), ageometrical-topological invariant, as a running example, I present differentuse cases that result in more efficient models for analyzing point clouds,graphs, and meshes. Moreover, I outline a vision for how topological conceptscould be used in the future, comprising (1) the learning of functions ontopological spaces, (2) the building of hybrid models that imbue neuralnetworks with knowledge about the topological information in data, and (3) theanalysis of qualitative properties of neural networks. With current researchalready addressing some of these aspects, this article thus serves as anintroduction and invitation to this nascent area of research.