John D. Cook • 8/15/2026

Compressing a Hadamard matrix

This article discusses the recent discovery of a new Hadamard matrix and focuses on efficient ways to store such matrices. It explains that while a naive representation uses n² bits, algorithmic methods like Sylvester's or Paley's construction can reduce storage to just parameters. For sporadic matrices, it highlights that the last row is determined by others, symmetric matrices need only n(n+1)/2 bits, and skew matrices can be stored in n(n−1)/2 bits. The post is relevant to computer science and information theory, particularly in data compression and matrix representation.

0 comments

#algorithms #Orthogonal Matrices #Hadamard Matrix