Abstract.In their seminal paper, Valiant, Skyum, Berkowitz and Rackoff proved that arithmetic circuits can be balanced. That is, they showed that for every arithmetic circuit Ф of size s and degree r, there exists an arithmetic circuit Ψ of size poly (r, s) and depth O (log(r) log(s)) computing the same polynomial. In the first part of this paper, we follow the proof of Valiant el al. and show that syntactically multilinear arithmetic circuits can be balanced. That is, we show that if Ф is syntactically multilinear, then so is Ψ.Recently, a super-polynomial separation between multilinear arithmetic formula and circuit size was shown. In the second part of this paper, we use the result of the first part to simplify the proof of this separation. That is, we construct a (simpler) polynomial f (x1, ... , xn) such that
Every multilinear arithmetic formula computing f is of size nΩ(log(n)).There exists a syntactically multilinear arithmetic circuit of size poly(n) and depth O(log2(n)) computing f.
[1]
Ran Raz,et al.
Multi-linear formulas for permanent and determinant are of super-polynomial size
,
2004,
STOC '04.
[2]
Laurent Hyafil,et al.
On the parallel evaluation of multivariate polynomials
,
1978,
SIAM J. Comput..
[3]
Ran Raz,et al.
Separation of Multilinear Circuit and Formula Size
,
2006,
Theory Comput..
[4]
Ran Raz,et al.
A Lower Bound for the Size of Syntactically Multilinear Arithmetic Circuits
,
2007,
48th Annual IEEE Symposium on Foundations of Computer Science (FOCS'07).
[5]
Noam Nisan,et al.
Lower bounds for non-commutative computation
,
1991,
STOC '91.
[6]
Noam Nisan,et al.
Lower bounds on arithmetic circuits via partial derivatives
,
2005,
computational complexity.
[7]
Leslie G. Valiant,et al.
Fast Parallel Computation of Polynomials Using Few Processors
,
1983,
SIAM J. Comput..