Separation Between Read-once Oblivious Algebraic Branching Programs (ROABPs) and Multilinear Depth-three Circuits

被引:2
作者
Kayal, Neeraj [1 ]
Nair, Vineet [2 ]
Saha, Chandan [2 ]
机构
[1] Microsoft Res, Vigyan 9,Lavelle Rd, Bengaluru 560001, Karnataka, India
[2] Indian Inst Sci, Bengaluru 560012, Karnataka, India
关键词
Multilinear depth-three circuits; read-once oblivious algebraic branching programs; evaluation dimension; skewed partial derivatives; expander graphs; iterated matrix multiplication; ARITHMETIC CIRCUITS; HITTING-SETS; LOWER BOUNDS; HARDNESS;
D O I
10.1145/3369928
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
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 2(Omega(n)) size. (2) Any multilinear depth-three circuit computing IMMn,d (the iterated matrix multiplication polynomial formed by multiplying d, n x n symbolic matrices) has n(Omega(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 n(Omega(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.
引用
收藏
页数:27
相关论文
共 44 条
  • [1] Agrawal M, 2005, LECT NOTES COMPUT SC, V3821, P92, DOI 10.1007/11590156_6
  • [2] HITTING-SETS FOR ROABP AND SUM OF SET-MULTILINEAR CIRCUITS
    Agrawal, Manindra
    Gurjar, Rohit
    Korwar, Arpita
    Saxena, Nitin
    [J]. SIAM JOURNAL ON COMPUTING, 2015, 44 (03) : 669 - 697
  • [3] Agrawal M, 2013, STOC'13: PROCEEDINGS OF THE 2013 ACM SYMPOSIUM ON THEORY OF COMPUTING, P321
  • [4] Alon N., 1985, J COMBIN THEORY SER
  • [5] Alon N., 1986, COMBINATORICA
  • [6] [Anonymous], 1970, Problems in analysis (Papers dedicated to Salomon Bochner, 1969), DOI DOI 10.1515/9781400869312-013
  • [7] BUSER P, 1982, ANN SCI ECOLE NORM S, V15, P213
  • [8] Introducing Theranostics Journal - From the Editor-in-Chief
    Chen, Xiaoyuan
    [J]. THERANOSTICS, 2011, 1 : 1 - 2
  • [9] PROBABILISTIC REMARK ON ALGEBRAIC PROGRAM TESTING
    DEMILLO, RA
    LIPTON, RJ
    [J]. INFORMATION PROCESSING LETTERS, 1978, 7 (04) : 193 - 195
  • [10] DODZIUK J, 1984, SIAM J COMPUT, V284, P787