一、报告题目

On Computing a Center Persistence Diagram（持续性图中心的求解计算）

二、报告人

朱滨海，美国蒙大拿州立大学计算机科学系教授

三、时间

2019年7月24日上午10:00–12:00

四、地点

软件园校区软件学院办公楼202会议室

五、摘要

Persistence diagram is a new tool from computational topology to capture the topological and geometric changes for large point clouds (or more complex objects). This talk first introduces the basics on persistence diagrams (e.g., the bottleneck distance between two diagrams). Then, we consider the center persistence diagram problem, i.e., one whose maximum bottleneck distance to m given diagrams is minimized . We show that, when m=2 diagrams are given, the problem is polynomially solvable. When m=3, we prove its NP-hardness (in fact, NP-hard to approximate within a factor of 2). Finally, we give a tight factor-2 approximation for the problem. No prior knowledge on topology is needed for this talk.