Seminar, Hosted by Institute for Computational and Experimental Research in Mathematics (ICERM), Brown University
Fast Construction of Hierarchically Low-Rank Matrices Using Randomized Sketching
Abstract
Sherry Xiaoye Li surveys randomized construction of hierarchically low-rank matrices, including H/H2, HODLR, HSS, and butterfly formats with different off-diagonal structures. Applications include integral equations, boundary elements, discretized PDEs, and statistical or machine-learning kernel matrices, using either iterative matrix-vector products or direct factorization and solves. Constructing these representations from an implicit dense operator is often the main cost. The talk offers a unified view of sketch distributions, sketch sizes, approximation error bounds, high-performance implementation, applications, and open questions.
Topics
- hierarchically low-rank matrices, search World Wide
- H/H2, search World Wide
- HODLR, search World Wide
- HSS, search World Wide
- butterfly formats, search World Wide
- off-diagonal structures, search World Wide
- sketch distributions, search World Wide
- approximation error bounds, search World Wide
- direct factorization, broader topic: factorization