Coded Matrix Computation: Numerical Stability, Partial Stragglers, and Sparse Input Matrices
Offered By: Simons Institute via YouTube
Course Description
Overview
Explore high-dimensional matrix computations in distributed computing environments through this 43-minute talk by Aditya Ramamoorthy from Iowa State University. Delve into the challenges of large-scale distributed computing clusters, focusing on the problem of "stragglers" and their impact on job execution time. Learn how coding theory concepts are being adapted to address these issues, enabling result recovery with a threshold of completed worker node tasks. Examine the intersection of application-driven coding theory and matrix computations, considering aspects such as numerical stability, partial stragglers, and sparse input matrices. Gain insights into innovative approaches for enhancing the efficiency and reliability of matrix computations in machine learning and scientific computing applications.
Syllabus
Coded matrix computation: numerical stability, partial stragglers and sparse input matrices
Taught by
Simons Institute
Related Courses
Machine Learning and Deep Learning Maths - Matrix and Vector OperationsThe AI University via YouTube Structure and Matrices in Julia Programming - Lecture 3
The Julia Programming Language via YouTube Sparse Matrices in Sparse Analysis - Anna Gilbert
Institute for Advanced Study via YouTube Practical Quantum Circuits for Block Encodings of Sparse Matrices
Institute for Pure & Applied Mathematics (IPAM) via YouTube C++ Compile-Time Sparse Matrices for Linear Algebra and Tracking Applications
CppNow via YouTube