Graphcode: Learning from multiparameter persistent homology using graph neural networks
Preprint
Abstract
We introduce graphcodes, a novel multi-scale summary of the topological properties of a dataset that is based on the well-established theory of persistent homology. Graphcodes handle datasets that are filtered along two real-valued scale parameters. Such multi-parameter topological summaries are usually based on complicated theoretical foundations and difficult to compute; in contrast, graphcodes yield an informative and interpretable summary and can be computed as efficient as one-parameter summaries. Moreover, a graphcode is simply an embedded graph and can therefore be readily integrated in machine learning pipelines using graph neural networks. We describe such a pipeline and demonstrate that graphcodes achieve better classification accuracy than state-of-the-art approaches on various datasets.Keywords
All Related Versions
This publication has 15 references indexed in Scilit:
- Compression for 2-parameter persistent homologyComputational Geometry, 2022
- A kernel for multi-parameter persistent homologyComputers & Graphics: X, 2019
- Combinatorial presentation of multidimensional persistent homologyJournal of Pure and Applied Algebra, 2017
- Phat – Persistent Homology Algorithms ToolboxJournal of Symbolic Computation, 2017
- Persistent homology for the quantitative prediction of fullerene stabilityJournal of Computational Chemistry, 2014
- Topology of viral evolutionProceedings of the National Academy of Sciences, 2013
- Topology based data analysis identifies a subgroup of breast cancers with a unique mutational profile and excellent survivalProceedings of the National Academy of Sciences, 2011
- The Theory of Multidimensional PersistenceDiscrete & Computational Geometry, 2009
- Unzerlegbare Darstellungen Imanuscripta mathematica, 1972
- Corrections and Supplementaries to My Paper concerning Krull-Remak-Schmidt’s TheoremNagoya Mathematical Journal, 1950