Large-Scale Matrix Operations

on YouTube ↗29m

" summary

Covers numerical methods and data representations for large-scale matrix and tensor computations used in scientific computing and machine learning, including approaches to mitigate the curse of dimensionality for tasks like solving linear systems, eigenvalue problems, and optimization. Discusses optimal matrix/tensor representation techniques plus the libraries and hardware designed to execute large-scale matrix operations efficiently.

AI-written summary · about

" chapters (16)

1. 0:00Introduction
2. 0:51Satellite image models
3. 2:40Linear algebra in ML
4. 3:23Matrix compute costs
5. 5:08Linear systems applications
6. 6:58Iterative matrix methods
7. 9:28Matrix decomposition
8. 12:01Convergence conditions
9. 13:47Conjugate gradient method
10. 17:20Matrices in deep learning
11. 18:22Dense matrix multiply
12. 20:11Low-rank approximation
13. 22:42Hardware utilization
14. 24:31SIMD and MKL
15. 26:39Band matrix example
16. 27:41References and summary

" related talks

• 32mRandomized Numerical Linear Algebra: OverviewPetros Drineas · Randomized Numerical Linear Algebra and Applications 2018
• 10mDale Smith, Data Scientist, NexidiaDale Smith · MLconf ATL 2015
• 32mA Primer (or Refresher) On Linear Algebra for Data ScienceRuben van de Geer · PyData London 2019
• 29mImplementing Randomized Matrix Algorithms in Parallel and Distributed EnvironmentsMichael Mahoney · Parallel and Distributed Algorithms for Inference and Optimization (Simons Institute, Oct. 21–24, 2013)
• 25mMultiplying really really big matricesPaul Leopardi · PyCon Australia 2014 Science & Data miniconf
• 01h 31mPast, Present and Future of Randomized Numerical Linear Algebra IIMichael Mahoney, Petros Drineas · Big Data Boot Camp 2013
• 35mA Crash Course in Applied Linear AlgebraPatrick Landreman · PyData New York City 2019
• 28mNext-generation frameworks for Large-scale Machine LearningAnima Anandkuma · Scale By The Bay 2019