Abstract
We show an exponential separation between two well-studied models of algebraic computation, namely, read-once oblivious algebraic branching programs (ROABPs) and multilinear depth-three circuits. In particular, we show the following:
(1) There exists an explicit n-variate polynomial computable by linear sized multilinear depth-three circuits (with only two product gates) such that every ROABP computing it requires 2Ω(n) size.
(2) Any multilinear depth-three circuit computing IMMn,d (the iterated matrix multiplication polynomial formed by multiplying d, n × n symbolic matrices) has nΩ(d) size. IMMn,d can be easily computed by a poly(n,d) sized ROABP.
(3) Further, the proof of (2) yields an exponential separation between multilinear depth-four and multilinear depth-three circuits: There is an explicit n-variate, degree d polynomial computable by a poly(n) sized multilinear depth-four circuit such that any multilinear depth-three circuit computing it has size nΩ(d). This improves upon the quasi-polynomial separation of Reference [36] between these two models.
The hard polynomial in (1) is constructed using a novel application of expander graphs in conjunction with the evaluation dimension measure [15, 33, 34, 36], while (2) is proved via a new adaptation of the dimension of the partial derivatives measure of Reference [32]. Our lower bounds hold over any field.
- Manindra Agrawal. 2005. Proving lower bounds via pseudo-random generators. In Proceedings of the 25th International Conference on Foundations of Software Technology and Theoretical Computer Science (FSTTCS’05) (Lecture Notes in Computer Science), Vol. 3821. Springer, 92--105.Google Scholar
Digital Library
- Manindra Agrawal, Rohit Gurjar, Arpita Korwar, and Nitin Saxena. 2015. Hitting-sets for ROABP and sum of set-multilinear circuits. SIAM J. Comput. 44, 3 (2015), 669--697.Google Scholar
Cross Ref
- Manindra Agrawal, Chandan Saha, and Nitin Saxena. 2013. Quasi-polynomial hitting-set for set-depth- formulas. In Proceedings of the Symposium on Theory of Computing Conference (STOC’13). ACM, 321--330.Google Scholar
Digital Library
- N. Alon. 1986. Eigenvalues and expanders. Combinatorica (1986).Google Scholar
- N. Alon and V. D. Milman. 1985. Isoperimetric inequalities for graphs and superconcentrators. J. Combin. Theory Ser. (1985).Google Scholar
- P. Buser. 1982. A note on the isopoermitric constant. Annu. Sci. Ecole Norm. 4, 15 (1982), 213--230.Google Scholar
- J. Cheeger. 1970. Lower bound for the smallest eigenvalue of the laplacian. Prob. Anal. (1970), 195--199.Google Scholar
- Xi Chen, Neeraj Kayal, and Avi Wigderson. 2011. Partial derivatives in arithmetic complexity and beyond. Found. Trends Theor. Comput. Sci. 6, 1--2 (2011), 1--138.Google Scholar
Digital Library
- Rafael Mendes de Oliveira, Amir Shpilka, and Ben lee Volk. 2016. Subexponential size hitting sets for bounded depth multilinear formulas. Comput. Complex. 25, 2 (2016), 455--505.Google Scholar
Digital Library
- Richard A. DeMillo and Richard J. Lipton. 1978. A probabilistic remark on algebraic program testing. Info. Process. Lett. 7, 4 (1978), 193--195.Google Scholar
Cross Ref
- J. Dodziuk. 1984. Difference equations, isoperimetric inequality and transience of certain random walks. SIAM J. Comput. 284, 2 (1984), 787--794.Google Scholar
- Zeev Dvir, Guillaume Malod, Sylvain Perifel, and Amir Yehudayoff. 2012. Separating multilinear branching programs and formulas. In Proceedings of the 44th Symposium on Theory of Computing Conference (STOC’12). 615--624.Google Scholar
Digital Library
- Zeev Dvir, Amir Shpilka, and Amir Yehudayoff. 2009. Hardness-randomness tradeoffs for bounded depth arithmetic circuits. SIAM J. Comput. 39, 4 (2009), 1279--1293.Google Scholar
Digital Library
- Michael A. Forbes, Ramprasad Saptharishi, and Amir Shpilka. 2014. Hitting sets for multilinear read-once algebraic branching programs, in any order. In Proceedings of the Symposium on Theory of Computing (STOC’14). ACM, 867--875.Google Scholar
Digital Library
- Michael A. Forbes and Amir Shpilka. 2013. Quasipolynomial-time identity testing of non-commutative and read-once oblivious algebraic branching programs. In Proceedings of the 54th Annual IEEE Symposium on Foundations of Computer Science (FOCS’13). 243--252.Google Scholar
- Hervé Fournier, Nutan Limaye, Guillaume Malod, and Srikanth Srinivasan. 2014. Lower bounds for depth-4 formulas computing iterated matrix multiplication. In Proceedings of the Symposium on Theory of Computing (STOC’14). 128--135.Google Scholar
Digital Library
- Ankit Gupta, Pritish Kamath, Neeraj Kayal, and Ramprasad Saptharishi. 2013. Arithmetic circuits: A chasm at depth three. In Proceedings of the 54th Annual IEEE Symposium on Foundations of Computer Science (FOCS’13). 578--587.Google Scholar
Digital Library
- Rohit Gurjar, Arpita Korwar, Nitin Saxena, and Thomas Thierauf. 2015. Deterministic identity testing for sum of read-once oblivious arithmetic branching programs. In Proceedings of the 30th Conference on Computational Complexity (CCC’15). 323--346.Google Scholar
Digital Library
- Philip Hall. 1935. On representatives of subsets. J. London Math. Soc. 10, 1 (1935), 26--30.Google Scholar
- Joos Heintz and Claus-Peter Schnorr. 1980. Testing polynomials which are easy to compute (extended abstract). In Proceedings of the 12th Annual ACM Symposium on Theory of Computing. ACM, 262--272.Google Scholar
Digital Library
- Shlomo Hoory, Nathan Linial, and Avi Wigderson. 2006. Expander graphs and their applications. Bull. Amer. Math. Soc. 43, 4 (2006), 439--561.Google Scholar
Cross Ref
- Stasys Jukna. 2015. Lower bounds for tropical circuits and dynamic programs. Theory Comput. Syst. 57, 1 (2015), 160--194.Google Scholar
Digital Library
- Valentine Kabanets and Russell Impagliazzo. 2004. Derandomizing polynomial identity tests means proving circuit lower bounds. Comput. Complex. 13, 1–2 (2004), 1--46.Google Scholar
Digital Library
- Erich Kaltofen. 1989. Factorization of polynomials given by straight-line programs. In Randomness and Computation. JAI Press, 375--412.Google Scholar
- Neeraj Kayal, Chandan Saha, and Sébastien Tavenas. 2015. Formulas having low individual degree.Google Scholar
- Neeraj Kayal and Ramprasad Saptharishi. 2014. A selection of lower bounds for arithmetic circuits. Perspect. Comput. Complex. (2014).Google Scholar
- Adam Klivans and Daniel A. Spielman. 2001. Randomness efficient identity testing of multivariate polynomials. In Proceedings on 33rd Annual ACM Symposium on Theory of Computing. 216--223.Google Scholar
- László Lovász. 1979. On determinants, matchings, and random algorithms. In Proceedings of the Symposium on Fundamentals of Computation Theory (FCT’79). 565--574.Google Scholar
- Meena Mahajan and V. Vinay. 1997. Determinant: Combinatorics, algorithms, and complexity. Chicago J. Theor. Comput. Sci. 1997 (1997).Google Scholar
Digital Library
- Noam Nisan. 1991. Lower bounds for non-commutative computation (extended abstract). In Proceedings of the 23rd Annual ACM Symposium on Theory of Computing. 410--418.Google Scholar
- Noam Nisan and Avi Wigderson. 1994. Hardness vs. randomness. J. Comput. Syst. Sci. 49, 2 (1994), 149--167.Google Scholar
Digital Library
- Noam Nisan and Avi Wigderson. 1997. Lower bounds on arithmetic circuits via partial derivatives. Comput. Complex. 6, 3 (1997), 217--234.Google Scholar
Digital Library
- Ran Raz. 2006. Separation of multilinear circuit and formula size. Theory Comput. 2, 1 (2006), 121--135.Google Scholar
Cross Ref
- Ran Raz. 2009. Multi-linear formulas for permanent and determinant are of super-polynomial size. J. ACM 56, 2 (2009).Google Scholar
Digital Library
- Ran Raz and Amir Yehudayoff. 2008. Balancing syntactically multilinear arithmetic circuits. Comput. Complex. 17, 4 (2008), 515--535.Google Scholar
Digital Library
- Ran Raz and Amir Yehudayoff. 2009. Lower bounds and separations for constant depth multilinear circuits. Comput. Complex. 18, 2 (2009), 171--207.Google Scholar
Cross Ref
- Ramprasad Saptharishi. 2014. Recent progress on arithmetic circuit lower bounds. Bull. EATCS 114 (2014).Google Scholar
- Nitin Saxena. 2009. Progress on polynomial identity testing. Bull. EATCS 99 (2009), 49--79.Google Scholar
- Nitin Saxena. 2013. Progress on polynomial identity testing II. Electr. Colloq. Comput. Complex. 20 (2013), 186.Google Scholar
- Jacob T. Schwartz. 1980. Fast probabilistic algorithms for verification of polynomial identities. J. ACM 27, 4 (1980), 701--717.Google Scholar
Digital Library
- Amir Shpilka and Amir Yehudayoff. 2010. Arithmetic circuits: A survey of recent results and open questions. Found. Trends Theor. Comput. Sci. 5, 3–4 (Mar. 2010), 207--388. DOI:https://doi.org/10.1561/0400000039Google Scholar
- Amir Shpilka and Amir Yehudayoff. 2010. Arithmetic circuits: A survey of recent results and open questions. Found. Trends Theor. Comput. Sci. 5, 3–4 (2010), 207--388.Google Scholar
- W. T. Tutte. 1947. The factorization of linear graphs. J. London Math. Soc. 22 (1947), 107--111.Google Scholar
Cross Ref
- L. G. Valiant. 1979. Completeness Classes in Algebra. In Proceedings of the 11th Annual ACM Symposium on Theory of Computing (STOC 79). ACM Press, New York, NY, 249--261.Google Scholar
Digital Library
- Richard Zippel. 1979. Probabilistic algorithms for sparse polynomials. In Proceedings of the International Symposium on Symbolic and Algebraic Computation (EUROSAM’79). 216--226.Google Scholar
Cross Ref
Index Terms
Separation Between Read-once Oblivious Algebraic Branching Programs (ROABPs) and Multilinear Depth-three Circuits
Recommendations
Reconstruction of Full Rank Algebraic Branching Programs
An algebraic branching program (ABP) A can be modelled as a product expression X1amp;middot; X2… Xd, where X1 and Xd are 1 × w and w × 1 matrices, respectively, and every other Xk is a w × w matrix; the entries of these matrices are linear forms in m ...
Deterministic identity testing for sum of read-once oblivious arithmetic branching programs
CCC '15: Proceedings of the 30th Conference on Computational ComplexityA read-once oblivious arithmetic branching program (ROABP) is an arithmetic branching program (ABP) where each variable occurs in at most one layer. We give the first polynomial time whitebox identity test for a polynomial computed by a sum of ...
On algebraic branching programs of small width
CCC '17: Proceedings of the 32nd Computational Complexity ConferenceIn 1979 Valiant showed that the complexity class VPe of families with polynomially bounded formula size is contained in the class VPs of families that have algebraic branching programs (ABPs) of polynomially bounded size. Motivated by the problem of ...






Comments