Reinforcement Learning Applied to Stochastic Dynamics on Persistence Diagram Space

arXiv 2608.06276·low signalXBlueskyLinkedInCopy link

This paper (2608.06276, submitted 2026-08-06 by Farzana Nasrin) observes that topological data analysis literature treats persistence diagrams as static objects, with limited frameworks for modeling how they evolve. It introduces stochastic dynamics on persistence diagram space driven by reinforcement learning, extending TDA from descriptive summaries toward controllable, evolving representations of multiscale topological structure. Narrow and theoretical, but relevant if you use topological features as model inputs on data that changes over time.