High-Performance Matrix-Based Graph Analysis

Explore cutting-edge approaches to high-performance graph analysis through this comprehensive minisymposium featuring four expert presentations on matrix-based computational methods for large-scale network analysis. Discover how numerical linear algebra foundations and decades of sparse matrix computation experience are being leveraged to tackle complex problems across social networks, biology, infrastructure, finance, and cybersecurity domains. Learn about distributed "hypersparse" matrices for cybersecurity applications in the opening presentation on modeling large-scale network traffic with anonymized data structures. Examine the SuiteSparse implementation of the GraphBLAS standard, which provides a mathematical framework for expressing graph algorithms using linear algebra operations. Gain insights into the new Sparse BLAS software API standard that standardizes interfaces for sparse linear algebra operations, enhancing portability and performance across different computing platforms. Understand Arachne, an innovative open-source interactive framework designed for real-world graph analytics problems, demonstrating practical applications of these theoretical concepts. Each presentation includes detailed technical discussions followed by a comprehensive Q&A session, providing opportunities to engage with leading researchers from MIT, Texas A&M University, and New Jersey Institute of Technology who are advancing the field of high-performance graph computation.