Multiparameter Persistence Based Losses in Machine Learning

This 14-minute tutorial from the Applied Algebraic Topology Network demonstrates how to incorporate persistence-based losses into machine learning pipelines. Learn through a simple 4-vertex simplicial complex example that explains the mathematical framework for using invariants from both single and multiparameter persistence in machine learning. Access the accompanying Python package and code examples at the provided link. The content is based on joint research with Luis Scoccola, David Loiseaux, Mathieu Carrière, and Steve Oudot, as published in "Differentiability and Optimization of Multiparameter Persistent Homology."