Student Brown Bag Seminar

Comparison of two scalar fields can be both difficult to define and difficult to implement. Topological data analysis has constructed tools to analyze such data such as persistence diagrams and merge trees. The former has well-defined and tractable distances which can determine the similarity. Merge trees, being a more complicated structure, inherently run into large computational hurdles in constructing distances which are robust to perturbation, but in theory provide more analytical power in many cases. We construct a distance on merge trees, called Merge Tree Matching distance and show an implementation of this distance using a A*. We experimentally show its analytical power as well as its computational bottlenecks.

Place: Math, 402 and Zoom:   https://arizona.zoom.us/j/82075792519  Password:  150721