High-Performance Matrix-Based Graph Analysis Minisymposium

Computation on networks and graphs at very large scale is important in applications as diverse as social networks, biology, infrastructure, finance, and cybersecurity. Much current work on high-performance graph analysis leans heavily on numerical linear algebra, both for mathematical foundations and for some 50 years of experience with high-performance sparse matrix computation.

This minisymposium will present a variety of perspectives relating high-performance computation on graphs and on matrices, focusing on practical tools and applications. The first talk highlights the use of distributed “hypersparse' matrices at very large scale in an application to cybersecurity. This is followed by a talk describing the SuiteSparse implementation of the GraphBLAS standard, which underpins that project among others. The final two talks focus respectively on matrices and on graphs: The third describes new work on the Sparse BLAS software API standard, and the fourth presents Arachne, an interactive system for large-scale graph analysis.

Join us for these 4 talks:
00:00 Modeling Large-Scale Network Traffic with Anonymized Hypersparse Matrices with Jeremy Kepner, Vijay Gadepally, and Hayden Jananthan or Massachusetts Institute of Technology, U.S.
24:00 Suitesparse:Graphblas: Graph Algorithms in the Language of Linear Algebra with Timothy A. Davis, Texas A&M University, U.S.
51:03 Sparseblas: The New Interface to Sparse Linear Algebra Operations with Piotr Luszczek, Massachusetts Institute of Technology, U.S.
1:19:03 Arachne: An Open-Source Interactive Graph Analytics Framework for Real-World Problems with David A. Bader, Oliver Alvarado Rodriguez, Zhihui Du, Fuhuan Li, and Mohammad Dindoost, New Jersey Institute of Technology, U.S.
1:41:50 Q&A

Learn more about SIAM https://www.siam.org/
Join SIAM https://www.siam.org/membership/indiv...
Attend a SIAM Conference https://www.siam.org/conferences-events