The geometry of rank decompositions of matrix multiplication II: 3 × 3 matrices

Abstract

This is the second in a series of papers on rank decompositions of the matrix multiplication tensor. We present new rank 23 decompositions for the 3×3 matrix multiplication tensor M ⟨3⟩ . All our decompositions have symmetry groups that include the standard cyclic permutation of factors but otherwise exhibit a range of behavior. One of them has 11 cubes as summands and admits an unexpected symmetry group of order 12.We establish basic information regarding symmetry groups of decompositions and outline two approaches for finding new rank decompositions of M ⟨n⟩ for larger n.PO Box 7311,

Probabilistic Tensors and Opportunistic Boolean Matrix Multiplication

“…The witnessing probability can be amplified from the value p for a single sketch to 1 − for any > 0 by repeating the sketch with independent randomness at least r = p −1 ln −1 times and taking the disjunction of the sketches obtained. 10 Setting = δ/(su) and using the union bound, we obtain that at least p −1 (ln δ −1 + ln su) repeated sketches recover the correct Boolean product with probability at least 1 − δ.…”

Strassen’s $$2 \times 2$$ matrix multiplication algorithm: a conceptual perspective

“…Several earlier publications note their importance [16,27]. The paper [2] explores symmetries of algorithms for 3 × 3 matrix multiplication.…”

The Geometry of Rank Decompositions of Matrix Multiplication I: 2 × 2 Matrices

“…In this paper we study Strassen's rank 7 decomposition of M 2 , which we denote Str. In the next paper [1] new decompositions of M 3 are presented and their symmetry groups are described. Although this project began before the papers [3,4] appeared, we have benefited greatly from them in our study.…”