Lab Home | Phone | Search
Center for Nonlinear Studies  Center for Nonlinear Studies
 Home 
 People 
 Current 
 Affiliates 
 Visitors 
 Students 
 Research 
 ICAM-LANL 
 Publications 
 Conferences 
 Workshops 
 Sponsorship 
 Talks 
 Colloquia 
 Colloquia Archive 
 Seminars 
 Postdoc Seminars Archive 
 Quantum Lunch 
 Quantum Lunch Archive 
 CMS Colloquia 
 Q-Mat Seminars 
 Q-Mat Seminars Archive 
 P/T Colloquia 
 Archive 
 Kac Lectures 
 Kac Fellows 
 Dist. Quant. Lecture 
 Ulam Scholar 
 Colloquia 
 
 Jobs 
 Postdocs 
 CNLS Fellowship Application 
 Students 
 Student Program 
 Visitors 
 Description 
 Past Visitors 
 Services 
 General 
 
 History of CNLS 
 
 Maps, Directions 
 CNLS Office 
 T-Division 
 LANL 
 
Wednesday, October 12, 2016
3:00 PM - 4:00 PM
CNLS Conference Room (TA-3, Bldg 1690)

Seminar

Topological Analysis at the Extreme Scale: Finding Features in Large Data Sets

Gunther Weber
Lawrence Berkeley National Laboratory

Scalar functions characterize many physical phenomena, e.g., the density distribution of particles highlights the structure of the universe in cosmological simulations, the fuel consumption rate indicates burning/extinct regions in combustion simulations, and the distance from a material boundary characterizes the structure of porous materials. Since data sizes in all applications grow beyond sizes that can be visualized directly, an increasingly important approach is to detect salient features in the data and use these features for subsequent analysis and visualization. Topological analysis has proven to be a powerful tool to detect such features for a wide range of applications since it supports feature definitions based on the general concept of “connectedness” and provides means to characterize the prominence and stability of features. However, the global nature of topological methods makes their parallelization challenging. In this talk, I will present applications of topological data analysis and a new approach to topological analysis on parallel machines: A new distributed representation of topological descriptors reduces the amount of communication necessary during their parallel computation and subsequent data analysis.

Host: Curt Canada