Topological Data Analysis Workshop

This talk fits into the general topic of 'stratification learning,' wherein one tries to make inferences about data based on some assumption that it is sampled from a mixture of manifolds glued together in some nicely-structured way. The theoretical tool of persistent local homology (PLH), now more than five years old, provides a useful way to understand the local singularity structure of the input dataset, where what one means by "local" can be thought of as a multi-scale input parameter.