Optimal Evaluation of Pairs of Bilinear Forms
Abstract
A large class of multiplication problems in arithmetic complexity can be viewed as the simultaneous evaluation of a set of bilinear forms. This class includes the multiplication of matrices, polynomials, quaternions, Cayley and complex numbers. Considering bilinear algorithms, the optimal number of nonscalar multiplications can be described as the rank of a three-tensor or as the smallest member of rank one matrices necessary to include a given set of matrices in their span.
In this paper, we attack a rather large subclass of three-tensors, namely that of $(p,q,2)$-tensors, for arbitrary p and q, and solve it completely in the case where the field of constants contains the roots of a polynomial associated with the given tensor. In all other cases, we prove that, in general, our bounds cannot be improved. The complexity of a general pair of bilinear forms is determined explicitly in terms of parameters related to Kronecker’s theory of pencils and to the theory of invariant polynomials. This reveals unexpected results and shows explicitly the dependence on the algebraic structure of the constants; we display, for example, a pair of $3 \times 3$ bilinear forms whose complexity is 3 over the field $Z_7 $ and which, however, requires exactly 4 nonscalar multiplications over the fields $Z_5 $ or $Z_{11} $. Corresponding optimal algorithms are described and several applications are considered.
Keywords
Get full access to this article
View all available purchase options and get full access to this article.
References
1.
Alfred V. Aho, John E. Hopcroft, Jeffrey D. Ullman, The design and analysis of computer algorithms, Addison-Wesley Publishing Co., Reading, Mass.-London-Amsterdam, 1975x+470
2.
M. D. Atkinson, N. M. Stephens, On the maximal multiplicative complexity of a family of bilinear forms, Linear Algebra Appl., 27 (1979), 1–8
3.
M. D. Atkinson, private communication
4.
Roger W. Brockett, David Dobkin, On the optimal evaluation of a set of bilinear forms, Linear Algebra and Appl., 19 (1978), 207–235
5.
Roger Brockett, David Dobkin, On the number of multiplications required for matrix multiplication, SIAM J. Comput., 5 (1976), 624–628
6.
D. Dobkin, Masters Thesis, On the arithmetic complexity of a class of arithmetic computations, Thesis, Harvard University, Cambridge, MA, 1973, September
7.
Charles M. Fiduccia, R. Miller, J. Thatcher, On obtaining upper bounds on the complexity of matrix multiplicationComplexity of computer computations (Proc. Sympos., IBM Thomas J. Watson Res. Center, Yorktown Heights, N.Y., 1972), Plenum, New York, 1972, 31–40, 187–212
8.
C. M. Fiduccia, Y. Zalcstein, Algebras having linear multiplicative complexities, Technical Report no. 46, Department of Computer Science, State University of New York at Stony Brook, 1975, August
9.
F. R. Gantmacher, The theory of matrices. Vols. 1, 2, Translated by K. A. Hirsch, Chelsea Publishing Co., New York, 1959Vol. 1, x+374 pp. Vol. 2, ix+276
10.
N. Gastinel, Sur le calcul des produits de matrices, Numer. Math., 17 (1971), 222–229
11.
N. Gastinel, Le Problèms De L'Extension Minimale Diagonals D'un Operateur Linéaire, no. 235, 1975, manuscript communicated by one of the referees
12.
D. Grigor'ev, N. K. Nikol'sku, On the algebraic complexity of computing a pair of bilinear formsInvestigations on Linear Operators, Izdat. “Nauka” Leningrad Otdel., Leningrad, 1974, 159–163
13.
J. E. Hopcroft, L. R. Kerr, On minimizing the number of multiplications necessary for matrix multiplication, SIAM J. Appl. Math., 20 (1971), 30–36
14.
J. Hopcroft, J. Musinski, Duality applied to the complexity of matrix multiplication and other bilinear forms, SIAM J. Comput., 2 (1973), 159–173
15.
T. D. Howell, Ph.D. Thesis, Tensor rank and the complexity of bilinear forms, Cornell University, 1976, Sept.
16.
T. D. Howell, J. C. Lafon, The complexity of the quaternionproduct, TR, 75-245, Department of Computer Science, Cornell University, 1975, June
17.
J. Ja' Ja', Ph.D. Thesis, On the algebraic complexity of classes of bilinear forms, Harvard University, 1977, Sept.
18.
J. Ja' Ja', Computation of bilinear forms over finite fields, Technical Report, CS-78-03, Department of Computer Science, Pennsylvania State University, 1978, Jan.
19.
Julian D. Laderman, A noncommutative algorithm for multiplying $3\times 3$ matrices using $23$ muliplications, Bull. Amer. Math. Soc., 82 (1976), 126–128, January
20.
Jean-Claude Lafon, Optimum computation of p bilinear forms, Linear Algebra and Appl., 10 (1975), 225–240
21.
Ju. V. Matijasevič, The Diophantineness of enumerable sets, Dokl. Akad. Nauk SSSR, 191 (1970), 279–282
22.
I. Munro, Problems related to matrix multiplication, Proceedings Courant Institute Symposium on Computational Complexity, New York, 1971, Oct.
23.
Morris Newman, Integral matrices, Academic Press, New York, 1972xvii+224
24.
R. Probert, On the complexity of symmetric computations, Technical Report, CS-73-02, University of Waterloo Computer Science, Waterloo, Ontario, 1973, Jan.
25.
Volker Strassen, Gaussian elimination is not optimal, Numer. Math., 13 (1969), 354–356
26.
Volker Strassen, R. Miller, J. Thatcher, Evaluation of rational functionsComplexity of computer computations (Proc. Sympos., IBM Thomas J. Watson Res. Center, Yorktown Heights, N.Y., 1972), Plenum, New York, 1972, 1–10, 187–212
27.
Volker Strassen, Vermeidung von Divisionen, J. Reine Angew. Math., 264 (1973), 184–202
28.
S. Winograd, On multiplication of $2\times 2$ matrices, Linear Algebra and Appl., 4 (1971), 381–388
Information & Authors
Information
Published In
Copyright
Copyright © 1979 Society for Industrial and Applied Mathematics.
History
Submitted: 8 November 1977
Published online: 13 July 2006
Keywords
Authors
Metrics & Citations
Metrics
Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.
Cited By
- An algorithm for the non-identifiability of rank-3 tensorsBollettino dell'Unione Matematica Italiana, Vol. 16, No. 3 | 15 March 2023
- Rank of a tensor and quantum entanglementLinear and Multilinear Algebra, Vol. 1 | 18 May 2023
- Maximal Rank of ${m}\times{n}\times 2$ Tensors over Arbitrary FieldsSIAM Journal on Matrix Analysis and Applications, Vol. 43, No. 2 | 6 June 2022
- Improved method for finding optimal formulas for bilinear maps in a finite fieldTheoretical Computer Science, Vol. 790 | 1 Oct 2019
- On the minimal ranks of matrix pencils and the existence of a best approximate block-term tensor decompositionLinear Algebra and its Applications, Vol. 561 | 1 Jan 2019
- Tensor rank is not multiplicative under the tensor productLinear Algebra and its Applications, Vol. 543 | 1 Apr 2018
- Low rank interpolation of boundary spline curvesComputer Aided Geometric Design, Vol. 55 | 1 Jul 2017
- Classification of subspaces in $${{\mathbb{F}}^2\otimes {\mathbb{F}}^3}$$ F 2 ⊗ F 3 and orbits in $${{\mathbb{F}}^2\otimes{\mathbb{F}}^3 \otimes {\mathbb{F}}^r}$$ F 2 ⊗ F 3 ⊗ F rJournal of Geometry, Vol. 108, No. 1 | 27 February 2016
- Typical ranks for 3-tensors, nonsingular bilinear maps and determinantal idealsJournal of Algebra, Vol. 471 | 1 Feb 2017
- A note on the ranks of 2 × 2 × 2 and 2 × 2 × 2 × 2 tensorsLinear and Multilinear Algebra, Vol. 61, No. 10 | 1 Oct 2013
- On Kruskal’s theorem that every 3 × 3 × 3 array has rank at most 5Linear Algebra and its Applications, Vol. 439, No. 2 | 1 Jul 2013
- Complexity Lower Bounds through Balanced Graph Properties2012 IEEE 27th Conference on Computational Complexity | 1 Jun 2012
- Subtracting a best rank-1 approximation from p × p × 2( p ≥2) tensorsNumerical Linear Algebra with Applications, Vol. 19, No. 3 | 7 April 2011
- Candecomp/Parafac: From Diverging Components to a Decomposition in Block TermsSIAM Journal on Matrix Analysis and Applications, Vol. 33, No. 2 | 17 April 2012
- Factorization strategies for third-order tensorsLinear Algebra and its Applications, Vol. 435, No. 3 | 1 Aug 2011
- Tensor Rank: Some Lower and Upper Bounds2011 IEEE 26th Annual Conference on Computational Complexity | 1 Jun 2011
- Subtracting a best rank-1 approximation may increase tensor rankLinear Algebra and its Applications, Vol. 433, No. 7 | 1 Dec 2010
- Rank of 3-tensors with 2 slices and Kronecker canonical formsLinear Algebra and its Applications, Vol. 431, No. 10 | 1 Oct 2009
- Tensor Decompositions and ApplicationsSIAM Review, Vol. 51, No. 3 | 5 August 2009
- A Method to Avoid Diverging Components in the Candecomp/Parafac Model for Generic $I\timesJ\times2$ ArraysSIAM Journal on Matrix Analysis and Applications, Vol. 30, No. 4 | 16 January 2009
- On the ideals of secant varieties to certain rational varietiesJournal of Algebra, Vol. 319, No. 5 | 1 Mar 2008
- Low-Rank Approximation of Generic $p \timesq \times2$ Arrays and Diverging Components in the Candecomp/Parafac ModelSIAM Journal on Matrix Analysis and Applications, Vol. 30, No. 3 | 25 September 2008
- Decompositions of a Higher-Order Tensor in Block Terms—Part III: Alternating Least Squares AlgorithmsSIAM Journal on Matrix Analysis and Applications, Vol. 30, No. 3 | 25 September 2008
- Degeneracy in Candecomp/Parafac explained for p × p × 2 arrays of rank p + 1 or higherPsychometrika, Vol. 71, No. 3 | 23 September 2006
- A Lower Bound on the Complexity of Polynomial Multiplication over Finite FieldsSIAM Journal on Computing, Vol. 34, No. 4 | 27 July 2006
- Ranks of tensors, secant varieties of Segre varieties and fat pointsLinear Algebra and its Applications, Vol. 355, No. 1-3 | 1 Nov 2002
- Maximal rank of an element of a tensor productLinear Algebra and its Applications, Vol. 182 | 1 Mar 1993
- The multiplicative complexity of quadratic boolean formsTheoretical Computer Science, Vol. 102, No. 2 | 1 Aug 1992
- Pencils of complex and real symmetric and skew matricesLinear Algebra and its Applications, Vol. 147 | 1 Mar 1991
- Maximal bilinear complexity and codesLinear Algebra and its Applications, Vol. 144 | 1 Jan 1991
- Multiplication of Polynomials over Finite FieldsSIAM Journal on Computing, Vol. 19, No. 3 | 13 July 2006
- Maximal Rank of $m\times n\times(mn-k)$ TensorsSIAM Journal on Computing, Vol. 19, No. 3 | 13 July 2006
- The multiplicative complexity of boolean functionsApplied Algebra, Algebraic Algorithms and Error-Correcting Codes | 1 June 2005
- Algebraic complexities and algebraic curves over finite fieldsJournal of Complexity, Vol. 4, No. 4 | 1 Dec 1988
- Multiplicative complexity of polynomial multiplication over finite fields28th Annual Symposium on Foundations of Computer Science (sfcs 1987) | 1 Oct 1987
- The multiplicative complexity of quadratic Boolean functions28th Annual Symposium on Foundations of Computer Science (sfcs 1987) | 1 Oct 1987
- Tensor Rank and Border Rank of Band Toeplitz MatricesSIAM Journal on Computing, Vol. 16, No. 2 | 13 July 2006
- On the Validity of the Direct Sum ConjectureSIAM Journal on Computing, Vol. 15, No. 4 | 31 July 2006
- Border rank of m×n×(mn−q) tensorsLinear Algebra and its Applications, Vol. 79 | 1 Jul 1986
- The Ranks of $m \times n \times (mn - 2)$ TensorsSIAM Journal on Computing, Vol. 12, No. 4 | 13 July 2006
- The computational complexity of a set of quadratic functionsJournal of Computer and System Sciences, Vol. 24, No. 2 | 1 Apr 1982
- Computing lower bounds on tensor rank over finite fieldsJournal of Computer and System Sciences, Vol. 24, No. 1 | 1 Feb 1982
- On the Complexity of Bilinear Forms with CommutativitySIAM Journal on Computing, Vol. 9, No. 4 | 13 July 2006
- Computations of Bilinear Forms over Finite FieldsJournal of the ACM, Vol. 27, No. 4 | 1 October 1980
- Border rank of a p×q×2 tensor and the optimal approximation of a pair of bilinear formsAutomata, Languages and Programming | 24 May 2005
- An Addendum to Kronecker’s Theory of PencilsSIAM Journal on Applied Mathematics, Vol. 37, No. 3 | 12 July 2006