Randomized Numerical Linear Algebra : A Perspective on the Field With an Eye to Software · HackerPapers discussions

Randomized Numerical Linear Algebra : A Perspective on the Field With an Eye to Software — mentioned by wuubuu

Fri, 18 Jul 2025 23:19:22 GMT

Randomized matrix sketching is one way to get at this (see https://arxiv.org/abs/2302.11474), the problem is hardware is heavily optimized for dense multiplies so what you save in flops doesn't translate to real runtime speeds ups.

Paper and discussion history. A mention is not an endorsement.

Randomized Numerical Linear Algebra : A Perspective on the Field With an Eye to Software — mentioned by jey

Tue, 24 Sep 2024 19:30:57 GMT

There's also a number of other algorithms for randomized numerical linear algebra ("RandNLA") that are especially useful when dealing with matrices that are too big to fit in memory, commonly occurring in "big data" applications. Here's a recent survey paper with an eye towards practical applications: https://arxiv.org/abs/2302.11474

Paper and discussion history. A mention is not an endorsement.