Abstract
The exponent ω of matrix multiplication is the infimum over all real numbers c such that for all ε > 0 there is an algorithm that multiplies n - n matrices using at most O(nc+ε) arithmetic operations over an arbitrary field. A trivial lower bound on ω is 2, and the best known upper bound until recently was ω < 2.376 achieved by Coppersmith and Winograd in 1987. There were two independent improvements on ω, one by Stothers in 2010 who showed that ω < 2.374, and one by myself that ultimately resulted in ω < 2.373. Here I discuss the road to these improvements and conclude with some open questions.
- A. V. Aho, J. E. Hopcroft, and J. Ullman. The design and analysis of computer algorithms. Addison-Wesley Longman Publishing Co., Boston, MA, 1974. Google ScholarDigital Library
- D. Bini. Relations between exact and approximate bilinear algorithms. applications. Calcolo, 17(1):87--97, 1980.Google ScholarCross Ref
- D. Bini, M. Capovani, F. Romani, and G. Lotti. O(n2.7799) complexity for n x n approximate matrix multiplication. Inf. Process. Lett., 8(5):234--235, 1979.Google ScholarCross Ref
- M. Bläser. A lower bound for the rank of matrix multiplication over arbitrary fields. In Proc. FOCS, pages 45--, 1999.Google Scholar
- H. Cohn and C. Umans. Fast matrix multiplication using coherent configurations. In Proc. SODA, 2013.Google ScholarCross Ref
- D. Coppersmith and S. Winograd. Matrix multiplication via arithmetic progressions. J. Symbolic Computation, 9(3):251--280, 1990. Google ScholarDigital Library
- J. M. Landsberg. New lower bounds for the rank of matrix multiplication. CoRR, abs/1206.1530, 2012.Google Scholar
- E. Mossel, R. O'Donnell, and R. A. Servedio. Learning juntas. In Proc. STOC, pages 206--212, 2003. Google ScholarDigital Library
- A. Schönhage. Partial and total matrix multiplication. SIAM J. Comput., 10(3):434--455, 1981.Google ScholarDigital Library
- A. Shoshan and U. Zwick. All pairs shortest paths in undirected graphs with integer weights. In Proc. FOCS, pages 605--614, 1999. Google ScholarDigital Library
- A. Shpilka. Lower bounds for matrix product. SIAM J. on Computing, 32(5):1185--1200, 2003. Google ScholarDigital Library
- A. Stothers. Ph.D. Thesis, U. Edinburgh, 2010.Google Scholar
- V. Strassen. Vermeidung von Divisionen. Crelle J. Reine Angew. Math., 1973(264):184--202, 1973.Google Scholar
- L. G. Valiant. General context-free recognition in less than cubic time. Journal of Computer and System Sciences, 10:308--315, 1975. Google ScholarDigital Library
- V. Vassilevska Williams. Multiplying matrices faster than Coppersmith-Winograd. In Proc. STOC, pages 887--898, 2012. Google ScholarDigital Library
- U. Zwick. All pairs shortest paths using bridging sets and rectangular matrix multiplication. JACM, 49(3):289--317, 2002. Google ScholarDigital Library
Index Terms
- Algorithms column: An overview of the recent progress on matrix multiplication by Virginia Vassilevska Williams
Recommendations
Scaling Matrices to Prescribed Row and Column Maxima
A nonnegative symmetric matrix $B$ has row maxima prescribed by a given vector $r$, if for each index $i$, the maximum entry in the $i$th row of $B$ equals $r_i$. This paper presents necessary and sufficient conditions so that for a given nonnegative ...
Randomized QR with Column Pivoting
The dominant contribution to communication complexity in factorizing a matrix using QR with column pivoting is due to column-norm updates that are required to process pivot decisions. We use randomized sampling to approximate this process which dramatically ...
Computing Row and Column Counts for Sparse QR and LU Factorization
AbstractWe present algorithms to determine the number of nonzeros in each row and column of the factors of a sparse matrix, for both the QR factorization and the LU factorization with partial pivoting. The algorithms use only the nonzero structure of the ...
Comments