Fourier Sparsity, Spectral Norm, and the Log-Rank Conjecture

We study Boolean functions with sparse Fourier spectrum or small spectral norm, and show their applications to the Log-rank Conjecture for XOR functions f(x ⊕ y) - a fairly large class of functions including well studied ones such as Equality and Hamming Distance. The rank of the communication matrix M<sub>f</sub> for such functions is exactly the Fourier sparsity of f. Let d = deg<sub>2</sub>(f) be the F<sub>2</sub>-degree of f and DCC(f · ⊕) stand for the deterministic communication complexity for f(x ⊕ y). We show that 1) DCC(f · ⊕) = O(2<sup>d2/2</sup> log<sup>d-</sup>2 ∥f̂∥<sub>1</sub>). In particular, the Log-rank conjecture holds for XOR functions with constant F<sub>2</sub>-degree. 2) D<sup>CC</sup>(f · ⊕) = O(d∥f̂∥<sub>1</sub>) = O(√(rank(M<sub>f</sub>))). This improves the (trivial) linear bound by nearly a quadratic factor. We obtain our results through a degree-reduction protocol based on a variant of polynomial rank, and actually conjecture that the communication cost of our protocol is at most log<sup>O(1)</sup> rank(M<sub>f</sub>). The above bounds are obtained from different analysis for the number of parity queries required to reduce f's F<sub>2</sub>-degree. Our bounds also hold for the parity decision tree complexity of f, a measure that is no less than the communication complexity. Along the way we also prove several structural results about Boolean functions with small Fourier sparsity ∥f̂∥<sub>0</sub> or spectral norm ∥f̂∥<sub>1</sub>, which could be of independent interest. For functions f with constant F<sub>2</sub>-degree, we show that: 1) f can be written as the summation of quasi-polynomially many indicator functions of subspaces with ±-signs, improving the previous doubly exponential upper bound by Green and Sanders; 2) being sparse in Fourier domain is polynomially equivalent to having a small parity decision tree complexity; and 3) f depends only on polylog∥f̂∥<sub>1</sub> linear functions of input variables. For functions f with small spectral norm, we show that: 1) there is an affine subspace of co dimension ∥f̂∥<sub>1</sub> on which f(x) is a constant, and 2) there is a parity decision ∥f̂∥<sub>1</sub>log∥f̂∥<sub>0</sub> for computing f.

[1]  Amir Shpilka,et al.  On the structure of boolean functions with small spectral norm , 2013, Electron. Colloquium Comput. Complex..

[2]  Troy Lee,et al.  Composition Theorems in Communication Complexity , 2010, ICALP.

[3]  Ben Green,et al.  The distribution of polynomials over finite fields, with applications to the Gowers norms , 2007, Contributions Discret. Math..

[4]  Bruno Codenotti,et al.  Spectral Analysis of Boolean Functions as a Graph Eigenvalue Problem , 1999, IEEE Trans. Computers.

[5]  Zhiqiang Zhang,et al.  Communication complexities of symmetric XOR functions , 2009, Quantum Inf. Comput..

[6]  Eyal Kushilevitz,et al.  Learning decision trees using the Fourier spectrum , 1991, STOC '91.

[7]  Andrei Kotlov Rank and chromatic number of a graph , 1997, J. Graph Theory.

[8]  Alexander Russell,et al.  An Entropic Proof of Chang's Inequality , 2014, SIAM J. Discret. Math..

[9]  Vince Grolmusz On the Power of Circuits with Gates of Low L1 Norms , 1997, Theor. Comput. Sci..

[10]  Michael E. Saks,et al.  Lattices, mobius functions and communications complexity , 1988, [Proceedings 1988] 29th Annual Symposium on Foundations of Computer Science.

[11]  Hamed Hatami,et al.  Spectral Norm of Symmetric Functions , 2012, APPROX-RANDOM.

[12]  Yang Li,et al.  Tight Bounds on Communication Complexity of Symmetric XOR Functions in One-Way and SMP Models , 2011, TAMC.

[13]  Kurt Mehlhorn,et al.  Las Vegas is better than determinism in VLSI and distributed computing (Extended Abstract) , 1982, STOC '82.

[14]  Ashley Montanaro,et al.  On the communication complexity of XOR functions , 2009, ArXiv.

[15]  Andris Ambainis,et al.  The Quantum Communication Complexity of Sampling , 2003, SIAM J. Comput..

[16]  Yang Liu,et al.  Quantum and randomized communication complexity of XOR functions in the SMP model , 2013, Electron. Colloquium Comput. Complex..

[17]  Mei-Chu Chang A polynomial bound in Freiman's theorem , 2002 .

[18]  Zhiqiang Zhang,et al.  On the parity complexity measures of Boolean functions , 2010, Theor. Comput. Sci..

[19]  Rocco A. Servedio,et al.  Testing Fourier Dimensionality and Sparsity , 2009, SIAM J. Comput..

[20]  Gatis Midrijanis Exact quantum query complexity for total Boolean functions , 2004, quant-ph/0403168.

[21]  L. Dickson Linear Groups, with an Exposition of the Galois Field Theory , 1958 .

[22]  W. T. Gowers,et al.  A New Proof of Szemerédi's Theorem for Arithmetic Progressions of Length Four , 1998 .

[23]  A. Razborov Communication Complexity , 2011 .

[24]  Noam Nisan,et al.  On rank vs. communication complexity , 1994, Proceedings 35th Annual Symposium on Foundations of Computer Science.

[25]  W. Beckner Inequalities in Fourier analysis , 1975 .

[26]  T. Sanders,et al.  Boolean Functions with small Spectral Norm , 2006, math/0605524.

[27]  Paul Valiant The Log-Rank Conjecture and low degree polynomials , 2004, Inf. Process. Lett..

[28]  Elad Haramaty,et al.  On the structure of cubic and quartic polynomials , 2009, STOC '10.

[29]  Zvi Galil,et al.  Lower bounds on communication complexity , 1984, STOC '84.

[30]  W. T. Gowers,et al.  A NEW PROOF OF SZEMER ´ EDI'S THEOREM , 2001 .

[31]  W. T. Gowers,et al.  A new proof of Szemerédi's theorem , 2001 .

[32]  H. Buhrman,et al.  Complexity measures and decision tree complexity: a survey , 2002, Theor. Comput. Sci..

[33]  Xiaoming Sun,et al.  Randomized Communication Complexity for Linear Algebra Problems over Finite Fields , 2012, STACS.

[34]  Shachar Lovett,et al.  Worst Case to Average Case Reductions for Polynomials , 2008, 2008 49th Annual IEEE Symposium on Foundations of Computer Science.

[35]  Shachar Lovett,et al.  An Additive Combinatorics Approach Relating Rank to Communication Complexity , 2012, 2012 IEEE 53rd Annual Symposium on Foundations of Computer Science.

[36]  László Lovász,et al.  The rank and size of graphs , 1996, J. Graph Theory.

[37]  Andrew Chi-Chih Yao,et al.  Some complexity questions related to distributive computing(Preliminary Report) , 1979, STOC.

[38]  Miklos Santha,et al.  Query Complexity of Matroids , 2013, CIAC.

[39]  Noga Alon,et al.  Testing Reed-Muller codes , 2005, IEEE Transactions on Information Theory.