Efficient approximations of matrix multiplication using truncated decompositions

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kar, Suvendu, M., Hariprasad, N., Sai Gowri J., Venkatapathi, Murugesan
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910161089593344
author Kar, Suvendu
M., Hariprasad
N., Sai Gowri J.
Venkatapathi, Murugesan
author_facet Kar, Suvendu
M., Hariprasad
N., Sai Gowri J.
Venkatapathi, Murugesan
contents We exploit the truncated singular value decomposition and the recently proposed circulant decomposition for an efficient first-order approximation of the multiplication of large dense matrices. A decomposition of each matrix into a sum of a sparse matrix with relatively few dominant entries and a dense residue can also use the above approach, and we present methods for multiplication using a Fourier decomposition and a cycle decomposition-based sparsifications. The proposed methods scale as $\mathcal{O}(n^2 \log n)$ in arithmetic operations for $n \times n$ matrices for usable tolerances in relative error $\sim$ 1\%. We also present demonstrations of large gains in the efficiency and speed of end-to-end operations of Large Language Models (LLMs) as a motivation. Note that different decompositions for the two matrices $A$ and $B$ in the product $AB$ are also possible in this approach, using efficient a priori evaluations for suitability, to improve further on the error tolerances demonstrated here.
format Preprint
id arxiv_https___arxiv_org_abs_2504_19308
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Efficient approximations of matrix multiplication using truncated decompositions
Kar, Suvendu
M., Hariprasad
N., Sai Gowri J.
Venkatapathi, Murugesan
Numerical Analysis
We exploit the truncated singular value decomposition and the recently proposed circulant decomposition for an efficient first-order approximation of the multiplication of large dense matrices. A decomposition of each matrix into a sum of a sparse matrix with relatively few dominant entries and a dense residue can also use the above approach, and we present methods for multiplication using a Fourier decomposition and a cycle decomposition-based sparsifications. The proposed methods scale as $\mathcal{O}(n^2 \log n)$ in arithmetic operations for $n \times n$ matrices for usable tolerances in relative error $\sim$ 1\%. We also present demonstrations of large gains in the efficiency and speed of end-to-end operations of Large Language Models (LLMs) as a motivation. Note that different decompositions for the two matrices $A$ and $B$ in the product $AB$ are also possible in this approach, using efficient a priori evaluations for suitability, to improve further on the error tolerances demonstrated here.
title Efficient approximations of matrix multiplication using truncated decompositions
topic Numerical Analysis
url https://arxiv.org/abs/2504.19308