Topological machine learning in a nutshell
If you are reading this, you are probably wondering about what topological machine learning can do for you and your projects. The purpose of this page is to provide a pithy and coarse, but hopefully useful, introduction to some concepts.
High-level View
If you are an expert in machine learning, you could summarise topological machine learning as novel set of inductive biases for models, making it possible to leverage properties such as connectivity of data set. To some extent, such properties are already covered by existing methods, but topology can be very useful as an additional ‘lens’ through which to view data. In graph learning tasks, for instance, being able to capture and use topological features—such as cycles—is crucial in order to improve predictive performance.
Additional Resources
Here are some additional resources that might be of interest:
Amézquita et al., “The Shape of Things to Come: Topological Data Analysis and Biology, from Molecules to Organisms”, Developmental Dynamics Volume 249, Issue 7, pp. 816–833, 2020. doi:10.1002/dvdy.175
Hensel et al., “A Survey of Topological Machine Learning Methods”, Frontiers in Artificial Intelligence. doi:10.3389/frai.2021.681108