Neural Persistence

Neural persistence explores the use of topological data analysis to understand and improve neural networks and other complex systems. Current research focuses on developing and comparing persistence-based kernels for classification tasks, leveraging persistent homology to model dynamic systems like swarms, and addressing limitations of existing neural persistence measures by incorporating network architecture into the analysis, such as through deep graph persistence. These methods offer new ways to analyze the structure and dynamics of complex data, potentially leading to more efficient and interpretable machine learning models and a deeper understanding of collective behavior in various systems.