ABSTRACT
We initiate a direction for proving lower bounds on the size of non-commutative arithmetic circuits. This direction is based on a connection between lower bounds on the size of non-commutative arithmetic circuits and a problem about commutative degree four polynomials, the classical sum-of-squares problem: find the smallest n such that there exists an identity (x12+x22+•• + xk2)• (y1^2+y22+•• + yk2)= f12+f22+ ... +fn2, where each fi = fi(X,Y) is bilinear in X={x1,... ,xk} and Y={y1,..., yk}. Over the complex numbers, we show that a sufficiently strong super-linear lower bound on n in, namely, n ≥ k1+ε with ε >0, implies an exponential lower bound on the size of arithmetic circuits computing the non-commutative permanent.
More generally, we consider such sum-of-squares identities for any M polynomial h(X,Y), namely: h(X,Y) = f12+f22+...+fn2.
Again, proving n ≥ k1+ε in for any explicit h over the complex numbers gives an exponential lower bound for the non-commutative permanent. Our proofs relies on several new structure theorems for non-commutative circuits, as well as a non-commutative analog of Valiant's completeness of the permanent.
We proceed to prove such super-linear bounds in some restricted cases. We prove that n ≥ Ω(k6/5) in (1), if f1,..., fn are required to have integer coefficients. Over the real numbers, we construct an explicit M polynomial h such that n in (2) must be at least Ω(k2). Unfortunately, these results do not imply circuit lower bounds. We also present other structural results about non-commutative arithmetic circuits. We show that any non-commutative circuit computing an ordered non-commutative polynomial can be efficiently transformed to a syntactically multilinear circuit computing that polynomial. The permanent, for example, is ordered. Hence, lower bounds on the size of syntactically multilinear circuits computing the permanent imply unrestricted non-commutative lower bounds. We also prove an exponential lower bound on the size of non-commutative syntactically multilinear circuit computing an explicit polynomial. This polynomial is, however, not ordered and an unrestricted circuit lower bound does not follow.
- A. Barvinok. A simple polynomial time algorithm to approximatethe permanent within a simply exponential factor. Random Structures and Algorithms 14(1), pages 29--61, 1999. Google ScholarDigital Library
- W. Baur and V. Strassen. The complexity of partial derivatives. Theoretical computer science (22), pages 317--330, 1983.Google Scholar
- P. Burgisser. Completeness and reduction in algebraic complexity theory. Springer-Verlag Berlin Heidelberg 2000.Google Scholar
- S. Chien and A. Sinclair. Algebras with polynomial identities and computing the determinant. SIAM Journal on Computing 37, pages 252--266, 2007. Google ScholarDigital Library
- S. Chien, L. Rasmussen and A. Sinclair. Clifford algebras an approximating the permanent. STOC '02, pages 222--231, 2002. Google ScholarDigital Library
- J. von zur Gathen. Algebraic complexity theory. Ann. Rev. Comp. Sci. (3), pages 317--347, 1988. Google ScholarDigital Library
- C. Godsil and I. Gutman. On the matching polynomial of a graph. Algebraic Methods in Graph Theory, pages 241--249, 1981.Google Scholar
- L. Hyafil. On the parallel evaluation of multivariate polynomials. SIAM J. Comput. 8(2), pages 120--123, 1979.Google Scholar
- P. Hrubes and A. Yehudayoff. Homogeneous formulas and symmetric polynomials. arXiv:0907.2621 Google ScholarDigital Library
- P. Hrubes, A. Wigderson and A. Yehudayoff. Relationless completeness and separations. To appear in CCC. Google ScholarDigital Library
- A. Hurwitz. Über die Komposition der quadratischen Formen von beliebigvielen Variabeln. Nach. Ges. der Wiss. Göttingen, pages 309--316, 1898.Google Scholar
- A. Hurwitz. Über die Komposition der quadratischen Formen. Math. Ann., 88, pages 1--25, 1923.Google ScholarCross Ref
- I. M. James. On the immersion problem for real projective spaces. Bull. Am. Math. Soc., 69, pages 231--238, 1967.Google ScholarCross Ref
- M. Jerrum, A. Sinclair and E. Vigoda. A polynomial-time approximation algorithm for the permanent of a matrix with nonnegative entries. J. ACM 51(4), pages 671--697, 2004. Google ScholarDigital Library
- S. Jukna Boolean function complexity: advances and frontiers. Book in preparation. Google ScholarDigital Library
- N. Karmarkar, R. Karp, R. Lipton, L. Lovasz and M. Luby. A Monte-Carlo algorithm for estimating the permanent. SIAM Journal on Computing 22(2), pages 284--293, 1993. Google ScholarDigital Library
- T. Kirkman. On pluquatemions, and horaoid products of sums of n squares. Philos.Mag. (ser. 3), 33, pages 447--459; 494--509, 1848.Google Scholar
- K. Y. Lam. Some new results on composition of quadratic forms. Inventiones Mathematicae., 1985.Google ScholarCross Ref
- T. Y. Lam and T. Smith. On Yuzvinsky's monomial pairings. Quart. J. Math. Oxford. (2), 44, pages 215--237, 1993.Google ScholarCross Ref
- K. Mulmuley. On P vs. NP, Geometric Complexity Theory, and the Riemann Hypothesis. Technical Report, Computer Science department, The University of Chicago, 2009.Google Scholar
- N. Nisan. Lower bounds for non-commutative computation. STOC '91, pages 410--418, 1991. Google ScholarDigital Library
- N. Nisan and A. Wigderson. Lower bounds on arithmetic circuits via partial derivatives. Computational Complexity, vol. 6, pages 217--234, 1996. Google ScholarDigital Library
- A. Pfister. Zur Darstellung definiter Funktionen als Summe von Quadraten. ph Inventiones Mathematicae., 1967.Google Scholar
- J. Radon. Lineare scharen orthogonalen matrizen. Abh. Math. Sem. Univ. Hamburg 1, pages 2--14, 1922.Google Scholar
- R. Raz. Multi-linear formulas for permanent and determinant are of super-polynomial size. Journal of the Association for Computing Machinery 56 (2), 2009. ıgnore Google ScholarDigital Library
- R. Raz. Elusive functions and lower bounds for arithmetic circuits. To appear in Theory of Computing. Google ScholarDigital Library
- R. Raz and A. Yehudayoff. Lower bounds and separation for constant depth multilinear circuits. Proceedings of Computational Complexity, pages 128--139, 2008. Google ScholarDigital Library
- R. Raz, A. Shpilka and A. Yehudayoff. A lower bound for the size of syntactically multilinear arithmetic circuits. SIAM Journal on Computing 38 (4), pages 1624--1647, 2008. Google ScholarDigital Library
- D. B. Shapiro. Composition of quadratic forms. W. de Gruyter Verlag, 2000.Google Scholar
- . Strassen. Die berechnungskomplexitat von elementarsymmetrischen funktionen und von interpolationskoeffizienten. Numerische Mathematik (20), pages 238--251, 1973.Google Scholar
- V. Strassen. Vermeidung von Divisionen. J. Reine Angew. Math. 264, pages 182--202, 1973.Google Scholar
- L. G. Valiant. Completeness classes in algebra. STOC '79, pages 249--261. Google ScholarDigital Library
- S. Winograd. On the number of multiplications needed to compute certain functions. Comm. on Pure and Appl. Math. (23), pages 165--179, 1970.Google Scholar
- P. Yiu. Sums of squares formulae with integer coefficients. Canad. Math. Bull., 30, pages 318--324, 1987.Google ScholarCross Ref
- P. Yiu. On the product of two sums of 16 squares as a sum of squares of integral bilinear forms. Quart. J. Math. Oxford. (2), 41, pages 463--500, 1990.Google ScholarCross Ref
- S. Yuzvinsky. A series of monomial pairings. phLinear and multilinear algebra, 15, pages 19--119, 1984.Google Scholar
Index Terms
- Non-commutative circuits and the sum-of-squares problem
Recommendations
Set-multilinear and non-commutative formula lower bounds for iterated matrix multiplication
STOC 2022: Proceedings of the 54th Annual ACM SIGACT Symposium on Theory of ComputingAn Algebraic Formula for a polynomial P∈ [x1,…,xN] is an algebraic expression for P(x1,…,xN) using variables, field constants, additions and multiplications. Such formulas capture an algebraic analog of the Boolean complexity class NC1. Proving lower ...
Multi-linear formulas for permanent and determinant are of super-polynomial size
An arithmetic formula is multilinear if the polynomial computed by each of its subformulas is multilinear. We prove that any multilinear arithmetic formula for the permanent or the determinant of an n × n matrix is of size super-polynomial in n. ...
Non-commutative formulas and frege lower bounds: a new characterization of propositional proofs
CCC '15: Proceedings of the 30th Conference on Computational ComplexityDoes every Boolean tautology have a short propositional-calculus proof? Here, a propositional-calculus (i.e. Frege) proof is any proof starting from a set of axioms and deriving new Boolean formulas using a fixed set of sound derivation rules. ...
Comments