Searcharxiv⌕ Search

arXiv subjects

Alex Samorodnitsky

Publications and source records attributed to Alex Samorodnitsky.

36 records · Page 2Linked to original sources
arXiv2015

The "Most informative boolean function" conjecture holds for high noise

We prove the "Most informative boolean function" conjecture of Courtade and Kumar for high noise $ε\ge 1/2 - δ$, for some absolute constant $δ> 0$. Namely, if $X$ is uniformly distributed in $\{0,1\}^n$ and $Y$ is obtained by flipping each coordinate of $X$ independently with probability $ε$, then, provided $ε\ge 1/2 - δ$, for any boolean function $f$ holds $I(f(X);Y) \le 1 - H(ε)$. This conjecture was previously known to hold only for balanced functions.

cs.IT↗
arXiv2014

On Coset Leader Graphs of LDPC Codes

Our main technical result is that, in the coset leader graph of a linear binary code of block length n, the metric balls spanned by constant-weight vectors grow exponentially slower than those in $\{0,1\}^n$. Following the approach of Friedman and Tillich (2006), we use this fact to improve on the first linear programming bound on the rate of LDPC codes, as the function of their minimal distance. This improvement, combined with the techniques of Ben-Haim and Lytsin (2006), improves the rate vs distance bounds for LDPC codes in a significant sub-range of relative distances.

cs.IT↗
arXiv2014

Bounds on the permanent and some applications

We give new lower and upper bounds on the permanent of a doubly stochastic matrix. Combined with previous work, this improves on the deterministic approximation factor for the permanent. We also give a combinatorial application of the lower bound, proving S. Friedland's "Asymptotic Lower Matching Conjecture" for the monomer-dimer problem.

math.CO↗
arXiv2012

An inequality for functions on the Hamming cube

We prove an inequality for functions on the discrete cube extending the edge-isoperimetric inequality for sets. This inequality turns out to be equivalent to the following claim about random walks on the cube: Subcubes maximize 'mean first exit time' among all subsets of the cube of the same cardinality.

math.CO↗
arXiv2011

Computing the partition function for perfect matchings in a hypergraph

Given non-negative weights w_S on the k-subsets S of a km-element set V, we consider the sum of the products w_{S_1} ... w_{S_m} for all partitions V = S_1 cup ... cup S_m into pairwise disjoint k-subsets S_i. When the weights w_S are positive and within a constant factor, fixed in advance, of each other, we present a simple polynomial time algorithm to approximate the sum within a polynomial in m factor. In the process, we obtain higher-dimensional versions of the van der Waerden and Bregman-Minc bounds for permanents. We also discuss applications to counting of perfect and nearly perfect matchings in hypergraphs.

math.CO↗
arXiv2010

Lower bounds for designs in symmetric spaces

A design is a finite set of points in a space on which every "simple" functions averages to its global mean. Illustrative examples of simple functions are low-degree polynomials on the Euclidean sphere or on the Hamming cube. We prove lower bounds on designs in spaces with a large group of symmetries. These spaces include globally symmetric Riemannian spaces (of any rank) and commutative association schemes with 1-transitive group of symmetries. Our bounds are, in general, implicit, relying on estimates on the spectral behavior of certain symmetry-invariant linear operators. They reduce to the first linear programming bound for designs in globally symmetric Riemannian spaces of rank 1 or in distance regular graphs. The proofs are different though, coming from viewpoint of abstract harmonic analysis in symmetric spaces. As a dividend we obtain the following geometric fact: a design is large because a union of "spherical caps" around its points "covers" the whole space.

math.CO↗
arXiv2008

Inverse Conjecture for the Gowers norm is false

Let $p$ be a fixed prime number, and $N$ be a large integer. The 'Inverse Conjecture for the Gowers norm' states that if the "$d$-th Gowers norm" of a function $f:\F_p^N \to \F_p$ is non-negligible, that is larger than a constant independent of $N$, then $f$ can be non-trivially approximated by a degree $d-1$ polynomial. The conjecture is known to hold for $d=2,3$ and for any prime $p$. In this paper we show the conjecture to be false for $p=2$ and for $d = 4$, by presenting an explicit function whose 4-th Gowers norm is non-negligible, but whose correlation any polynomial of degree 3 is exponentially small. Essentially the same result (with different correlation bounds) was independently obtained by Green and Tao \cite{gt07}. Their analysis uses a modification of a Ramsey-type argument of Alon and Beigel \cite{ab} to show inapproximability of certain functions by low-degree polynomials. We observe that a combination of our results with the argument of Alon and Beigel implies the inverse conjecture to be false for any prime $p$, for $d = p^2$.

math.CO↗
arXiv2008

A modified logarithmic Sobolev inequality for the Hamming cube and some applications

The logarithmic Sobolev inequality for the Hamming cube {0,1}^n states that for any real-valued function f on the cube holds E(f,f) \ge 2 Ent(f^2), where E(f,f) is the appropriate Dirichlet form (also known as "sum of influences"). We show that the constant C = 2 at the right hand side of this inequality can be replaced by a function C(rho) depending on rho = Ent(f^2) / (n Ef^2). The function C is an increasing convex function taking [0,log 2] to [2, 2/log 2]. We present some applications of this modified inequality. In particular, it is used to obtain a discrete version of the Faber-Krahn inequality for small subsets of the Hamming cube, answering a question of Friedman and Tillich. We introduce, following the approach of Friedman and Tillich, the notion of a fractional edge-boundary size of a subset of {0,1}^n, and show Hamming balls of radius at most n/2 - O(n^{3/4}) to be sets with (asymptotically) the smallest fractional edge-boundary for their size.

math.CO↗
arXiv2008

An approximation algorithm for counting contingency tables

We present a randomized approximation algorithm for counting contingency tables, mxn non-negative integer matrices with given row sums R=(r_1, ..., r_m) and column sums C=(c_1, ..., c_n). We define smooth margins (R,C) in terms of the typical table and prove that for such margins the algorithm has quasi-polynomial N^{O(ln N)} complexity, where N=r_1+...+r_m=c_1+...+c_n. Various classes of margins are smooth, e.g., when m=O(n), n=O(m) and the ratios between the largest and the smallest row sums as well as between the largest and the smallest column sums are strictly smaller than the golden ratio (1+sqrt{5})/2 = 1.618. The algorithm builds on Monte Carlo integration and sampling algorithms for log-concave densities, the matrix scaling algorithm, the permanent approximation algorithm, and an integral representation for the number of contingency tables.

math.CO↗
arXiv2007

Counting magic squares in quasi-polynomial time

We present a randomized algorithm, which, given positive integers n and t and a real number 0< epsilon <1, computes the number Sigma(n, t) of n x n non-negative integer matrices (magic squares) with the row and column sums equal to t within relative error epsilon. The computational complexity of the algorithm is polynomial in 1/epsilon and quasi-polynomial in N=nt, that is, of the order N^{log N}. A simplified version of the algorithm works in time polynomial in 1/epsilon and N and estimates Sigma(n,t) within a factor of N^{log N}. This simplified version has been implemented. We present results of the implementation, state some conjectures, and discuss possible generalizations.

math.CO↗
arXiv2007

Linear programming bounds for codes via a covering argument

We recover the first linear programming bound of McEliece, Rodemich, Rumsey, and Welch for binary error-correcting codes and designs via a covering argument. It is possible to show, interpreting the following notions appropriately, that if a code has a large distance, then its dual has a small covering radius and, therefore, is large. This implies the original code to be small. We also point out (in conjunction with further work) that this bound is a natural isoperimetric constant of the Hamming cube, related to its Faber-Krahn minima. While our approach belongs to the general framework of Delsarte's linear programming method, its main technical ingredient is Fourier duality for the Hamming cube. In particular, we do not deal directly with Delsarte's linear program or orthogonal polynomial theory.

math.CO↗
arXiv2006

An upper bound for permanents of nonnegative matrices

A recent conjecture of Caputo, Carlen, Lieb, and Loss, and, independently, of the author, states that the maximum of the permanent of a matrix whose rows are unit vectors in l_p is attained either for the identity matrix I or for a constant multiple of the all-1 matrix J. The conjecture is known to be true for p = 1 (I) and for p \ge 2 (J). We prove the conjecture for a subinterval of (1,2), and show the conjectured upper bound to be true within a subexponential factor (in the dimension) for all 1 < p < 2. In fact, for p bounded away from 1, the conjectured upper bound is true within a constant factor. This leads to a mild (subexponential) improvement in deterministic approximation factor for the permanent. We present an efficient deterministic algorithm that approximates the permanent of a nonnegative n\times n matrix within exp{n - O(n/\log n)}.

math.CO↗
arXiv2006

Low-degree tests at large distances

We define tests of boolean functions which distinguish between linear (or quadratic) polynomials, and functions which are very far, in an appropriate sense, from these polynomials. The tests have optimal or nearly optimal trade-offs between soundness and the number of queries. In particular, we show that functions with small Gowers uniformity norms behave ``randomly'' with respect to hypergraph linearity tests. A central step in our analysis of quadraticity tests is the proof of an inverse theorem for the third Gowers uniformity norm of boolean functions. The last result has also a coding theory application. It is possible to estimate efficiently the distance from the second-order Reed-Muller code on inputs lying far beyond its list-decoding radius.

math.CO↗
arXiv2005

Gowers Uniformity, Influence of Variables, and PCPs

Gowers introduced, for d\geq 1, the notion of dimension-d uniformity U^d(f) of a function f: G -> \C, where G is a finite abelian group and \C are the complex numbers. Roughly speaking, if U^d(f) is small, then f has certain "pseudorandomness" properties. We prove the following property of functions with large U^d(f). Write G=G_1 x >... x G_n as a product of groups. If a bounded balanced function f:G_1 x ... x G_n -> \C is such that U^{d} (f) > epsilon, then one of the coordinates of f has influence at least epsilon/2^{O(d)}. The Gowers inner product of a collection of functions is a related notion of pseudorandomness. We prove that if a collection of bounded functions has large Gowers inner product, and at least one function in the collection is balanced, then there is a variable that has high influence for at least four of the functions in the collection. Finally, we relate the acceptance probability of the "hypergraph long-code test" proposed by Samorodnitsky and Trevisan to the Gowers inner product of the functions being tested and we deduce applications to the construction of Probabilistically Checkable Proofs and to hardness of approximation.

math.CO↗
arXiv2003

Random Weighting, Asymptotic Counting, and Inverse Isoperimetry

For a family X of k-subsets of the set 1,...,n, let |X| be the cardinality of X and let Gamma(X,mu) be the expected maximum weight of a subset from X when the weights of 1,...,n are chosen independently at random from a symmetric probability distribution mu on R. We consider the inverse isoperimetric problem of finding mu for which Gamma(X,mu) gives the best estimate of ln|X|. We prove that the optimal choice of mu is the logistic distribution, in which case Gamma(X,mu) provides an asymptotically tight estimate of ln|X| as k^{-1}ln|X| grows. Since in many important cases Gamma(X,mu) can be easily computed, we obtain computationally efficient approximation algorithms for a variety of counting problems. Given mu, we describe families X of a given cardinality with the minimum value of Gamma(X,mu), thus extending and sharpening various isoperimetric inequalities in the Boolean cube.

math.CO↗
arXiv2002

On linear programming bounds for spherical codes and designs

We investigate universal bounds on spherical codes and spherical designs that could be obtained using Delsarte's linear programming methods. We give a lower estimate for the LP upper bound on codes, and an upper estimate for the LP lower bound on designs. Specifically, when the distance of the code is fixed and the dimension goes to infinity, the LP upper bound on codes is at least as large as the average of the best known upper and lower bounds. When the dimension n of the design is fixed, and the strength k goes to infinity, the LP bound on designs turns out, in conjunction with known lower bounds, to be proportional to k^{n-1}.

math.CO↗
arXiv2000

The Distance Approach to Approximate Combinatorial Counting

We develop general methods to obtain fast (polynomial time) estimates of the cardinality of a combinatorially defined set via solving some randomly generated optimization problems on the set. Geometrically, we estimate the cardinality of a subset of the Boolean cube via the average distance from a point in the cube to the subset. As an application, we present a new randomized polynomial time algorithm which approximates the permanent of a 0-1 matrix by solving a small number of Assignment problems.

math.CO↗