We study the kernel complexity of constraint satisfaction problems over a finite domain, parameterized by the number of variables, whose constraint language consists of two relations: the non-equality relation and an additional permutation-invariant relation R. We establish a conditional lower bound on the kernel size in terms of the largest arity of an or relation definable from R. Building on this, we investigate the kernel complexity of uniformly rainbow free coloring problems. In these problems, for fixed positive integers d, 𝓁, and q ≥ d, we are given a graph G on n vertices and a collection F of 𝓁-tuples of d-subsets of its vertex set, and the goal is to decide whether there exists a proper coloring of G with q colors such that no 𝓁-tuple in F is uniformly rainbow, that is, no tuple has all its sets colored with the same d distinct colors. We determine, for all admissible values of d, 𝓁, and q, the infimum over all values η for which the problem admits a kernel of size O(n^η), under the assumption NP ⊈ coNP/poly. As applications, we obtain nearly tight bounds on the kernel complexity of various coloring problems under diverse settings and parameterizations. This includes graph coloring problems parameterized by the vertex-deletion distance to a disjoint union of cliques, resolving a question of Schalken (2020), as well as uniform hypergraph coloring problems parameterized by the number of vertices, extending results of Jansen and Pieterse (2019) and Beukers (2021).

A family of permutations of [n] is called setwise distinguishable if for every permutation in the family there exists a subset of [n] whose image under this permutation differs from its image under any other permutation in the family. We prove that there exists a setwise distinguishable family of 2^{(2-o(1))⋅n} permutations of [n]. The result is optimal up to the o(1) term in the exponent and is achieved through an explicit construction. As an application, we obtain nearly tight conditional lower bounds on the kernelization complexity of graph coloring problems parameterized by the vertex-deletion distance to split graphs. This improves a result of Jansen and Kratsch (Inf. Comput., 2013).

For a fixed graph H, the H-Coloring problem asks whether a given graph admits an edge-preserving function from its vertex set to that of H. A seminal theorem of Hell and Nešetřil asserts that the H-Coloring problem is NP-hard whenever H is loopless and non-bipartite. A result of Jansen and Pieterse implies that for every graph H, the H-Coloring problem parameterized by the vertex cover number k admits a kernel with O(k^Δ(H)) vertices and bit-size bounded by O(k^Δ(H)⋅log k), where Δ(H) denotes the maximum degree in H. For the case where H is a complete graph on at least three vertices, this kernel size nearly matches conditional lower bounds established by Jansen and Kratsch and by Jansen and Pieterse. This paper presents new upper and lower bounds on the kernel size of H-Coloring problems parameterized by the vertex cover number. The upper bounds arise from two kernelization algorithms. The first is purely combinatorial, and its size is governed by a structural quantity of the graph H, called the non-adjacency witness number. As applications, we obtain kernels whose size is bounded by a fixed polynomial for natural classes of graphs H with unbounded maximum degree, such as planar graphs and, more broadly, graphs with bounded degeneracy. More strikingly, we show that for almost every graph H, the degree of the polynomial that bounds the size of our combinatorial kernel grows only logarithmically in Δ(H). Our second kernel leverages linear-algebraic tools and involves the notion of faithful independent representations of graphs. It strengthens the general bound from prior work and, among other applications, yields near-optimal kernels for problems concerning the dimension of orthogonal graph representations over finite fields. We complement our kernelization results with conditional lower bounds, thereby nearly settling the kernel complexity of the problem for various target graphs H.

The low-rank matrix completion problem asks whether a given real matrix with missing values can be completed so that the resulting matrix has low rank or is close to a low-rank matrix. The completed matrix is often required to satisfy additional structural constraints, such as positive semi-definiteness or a bounded infinity norm. The problem arises in various research fields, including machine learning, statistics, and theoretical computer science, and has broad real-world applications. This paper presents new NP-hardness results for low-rank matrix completion problems. We show that for every sufficiently large integer d and any real number ε ∈ [2^{-O(d)},1/7], given a partial matrix A with exposed values of magnitude at most 1 that admits a positive semi-definite completion of rank d, it is NP-hard to find a positive semi-definite matrix that agrees with each given value of A up to an additive error of at most ε, even when the rank is allowed to exceed d by a multiplicative factor of O (1/(ε²⋅log(1/ε))). This strengthens a result of Hardt, Meka, Raghavendra, and Weitz (COLT, 2014), which applies to multiplicative factors smaller than 2 and to ε that decays polynomially in d. We establish similar NP-hardness results for the case where the completed matrix is constrained to have a bounded infinity norm (rather than be positive semi-definite), for which all previous hardness results rely on complexity assumptions related to the Unique Games Conjecture. Our proofs involve a novel notion of nearly orthonormal representations of graphs, the concept of line digraphs, and bounds on the rank of perturbed identity matrices.

A set family F is called intersecting if every two members of F intersect, and it is called uniform if all members of F share a common size. A uniform family F ⊆ binom([n],k) of k-subsets of [n] is ε-far from intersecting if one has to remove more than ε ⋅ binom(n,k) of the sets of F to make it intersecting. We study the property testing problem that given query access to a uniform family F ⊆ binom([n],k), asks to distinguish between the case that F is intersecting and the case that it is ε-far from intersecting. We prove that for every fixed integer r, the problem admits a non-adaptive two-sided error tester with query complexity O((ln n)/ε) for ε ≥ Ω((k/n)^r) and a non-adaptive one-sided error tester with query complexity O((ln k)/ε) for ε ≥ Ω((k²/n)^r). The query complexities are optimal up to the logarithmic terms. For ε ≥ Ω((k²/n)²), we further provide a non-adaptive one-sided error tester with optimal query complexity of O(1/ε). Our findings show that the query complexity of the problem behaves differently from that of testing intersectingness of non-uniform families, studied recently by Chen, De, Li, Nadimpalli, and Servedio (ITCS, 2024).

For a field $\mathbb{F}$ and integers $d$ and $k$, a set of vectors of $\mathbb{F}^d$ is called $k$-nearly orthogonal if its members are non-self-orthogonal and every $k+1$ of them include an orthogonal pair. We prove that for every prime $p$ there exists a positive constant $δ= δ(p)$, such that for every field $\mathbb{F}$ of characteristic $p$ and for all integers $k \geq 2$ and $d \geq k^{1/(p-1)}$, there exists a $k$-nearly orthogonal set of at least $d^{δ\cdot k^{1/(p-1)}/ \log k}$ vectors of $\mathbb{F}^d$. In particular, for the binary field we obtain a set of $d^{Ω( k /\log k)}$ vectors, and this is tight up to the $\log k$ term in the exponent. For comparison, the best known lower bound over the reals is $d^{Ω( \log k / \log \log k)}$ (Alon and Szegedy, Graphs and Combin., 1999). The proof combines probabilistic and spectral arguments.

We show that the Consensus Division theorem implies lower bounds on the chromatic number of Kneser hypergraphs, offering a novel proof for a result of Alon, Frankl, and Lovász (Trans. Amer. Math. Soc., 1986) and for its generalization by Kř\'ıž (Trans. Amer. Math. Soc., 1992). Our approach is applied to study the computational complexity of the total search problem Kneser$^p$, which given a succinct representation of a coloring of a $p$-uniform Kneser hypergraph with fewer colors than its chromatic number, asks to find a monochromatic hyperedge. We prove that for every prime $p$, the Kneser$^p$ problem with an extended access to the input coloring is efficiently reducible to a quite weak approximation of the Consensus Division problem with $p$ shares. In particular, for $p=2$, the problem is efficiently reducible to any non-trivial approximation of the Consensus Halving problem on normalized monotone functions. We further show that for every prime $p$, the Kneser$^p$ problem lies in the complexity class $\mathsf{PPA}$-$p$. As an application, we establish limitations on the complexity of the Kneser$^p$ problem, restricted to colorings with a bounded number of colors.

The orthogonality dimension of a graph over $\mathbb{R}$ is the smallest integer $d$ for which one can assign to every vertex a nonzero vector in $\mathbb{R}^d$ such that every two adjacent vertices receive orthogonal vectors. For an integer $d$, the $d$-Ortho-Dim$_\mathbb{R}$ problem asks to decide whether the orthogonality dimension of a given graph over $\mathbb{R}$ is at most $d$. We prove that for every integer $d \geq 3$, the $d$-Ortho-Dim$_\mathbb{R}$ problem parameterized by the vertex cover number $k$ admits a kernel with $O(k^{d-1})$ vertices and bit-size $O(k^{d-1} \cdot \log k)$. We complement this result by a nearly matching lower bound, showing that for any $\varepsilon > 0$, the problem admits no kernel of bit-size $O(k^{d-1-\varepsilon})$ unless $\mathsf{NP} \subseteq \mathsf{coNP/poly}$. We further study the kernelizability of orthogonality dimension problems in additional settings, including over general fields and under various structural parameterizations.

Abstract. The Kneser graph [Formula: see text] is defined for integers [Formula: see text] and [Formula: see text] with [Formula: see text] as the graph whose vertices are all the [Formula: see text]-subsets of [Formula: see text] where two such sets are adjacent if they are disjoint. The Schrijver graph [Formula: see text] is defined as the subgraph of [Formula: see text] induced by the collection of all [Formula: see text]-subsets of [Formula: see text] that do not include two consecutive elements modulo [Formula: see text]. It is known that the chromatic number of both [Formula: see text] and [Formula: see text] is [Formula: see text]. In the computational Kneser and Schrijver problems, we are given access to a coloring with [Formula: see text] colors of the vertices of [Formula: see text] and [Formula: see text], respectively, and the goal is to find a monochromatic edge. We prove that the problems admit randomized algorithms with running time [Formula: see text], hence they are fixed-parameter tractable with respect to the parameter [Formula: see text]. The analysis involves structural results on intersecting families and on induced subgraphs of Kneser and Schrijver graphs. We also study the Agreeable-Set problem of assigning a small subset of a set of [Formula: see text] items to a group of [Formula: see text] agents, so that all agents value the subset at least as much as its complement. As an application of our algorithm for the Kneser problem, we obtain a randomized polynomial-time algorithm for the Agreeable-Set problem for instances with [Formula: see text]. We further show that the Agreeable-Set problem is at least as hard as a variant of the Kneser problem with extended access to the input coloring.

The index coding problem is concerned with the amount of information that a sender has to transmit to multiple receivers in a way that enables each of them to retrieve its requested data relying on prior side information. For linear index coding, the problem is characterized by the minrank parameter of a graph that represents the side information map of the receivers. Previous work has shown that it is$\mathsf {NP}$-hard to determine the minrank parameter of graphs. In this work, we study the computational complexity of the minrank parameter on perturbed instances, obtained from worst-case instances by a random extension of the side information available to the receivers. This setting is motivated by applications of index coding, in which the side information is accumulated via repeated transmissions that suffer from loss of data due to noisy communication or storage capacity. We prove that determining the minrank parameter remains computationally hard on perturbed instances. Our contribution includes an extension of several hardness results of the minrank parameter to the perturbed setting as well as a general technique for deriving the hardness of the minrank parameter on perturbed instances from its hardness on worst-case instances.

The index coding problem is concerned with broadcasting encoded information to a collection of receivers in a way that enables each receiver to discover its required data based on its side information, which comprises the data required by some of the others. Given the side information map, represented by a graph in the symmetric case and by a digraph otherwise, the goal is to devise a coding scheme of minimum broadcast length. We present a general method for developing efficient algorithms for approximating the index coding rate for prescribed families of instances. As applications, we obtain polynomial-time algorithms that approximate the index coding rate of graphs and digraphs on n vertices to within factors of$O(n/\log ^{2} n)$and$O(n/\log n)$respectively. This improves on the approximation factors of$O(n/\log n)$for graphs and$O(n \cdot \log \log n/\log n)$for digraphs achieved by Blasiak, Kleinberg, and Lubetzky. For the family of quasi-line graphs, we exhibit a polynomial-time algorithm that approximates the index coding rate to within a factor of 2. This improves on the approximation factor of$O(n^{2/3})$achieved by Arbabjolfaei and Kim for graphs on n vertices taken from certain sub-families of quasi-line graphs. Our approach is applicable for approximating a variety of additional graph and digraph quantities to within the same approximation factors. Specifically, it captures every graph quantity sandwiched between the independence number and the clique cover number and every digraph quantity sandwiched between the maximum size of an acyclic induced sub-digraph and the directed clique cover number.

This paper focuses on the study of the minimum possible alphabet size of codes in a generalized setting where the coding scheme is required to handle a pre-specified set of erasure or error patterns, naturally represented by a hypergraph. The need for such codes arises in many settings of practical interest, including wireless communication and flash memory systems. In many such settings, a smaller field size is achievable than that offered by MDS and other standard codes. We establish a connection between the minimum alphabet size of codes in this generalized setting and the combinatorial properties of the hypergraph that represents the pre-specified collection of erasure or error patterns. We also establish connections between error and erasure correcting codes in our generalized setting. Finally, we consider a variation of the problem that allows a small probability of decoding error and relate it to an approximate version of hypergraph coloring.

The orthogonality dimension of a graph G over ℝ is the smallest integer k for which one can assign a nonzero k-dimensional real vector to each vertex of G, such that every two adjacent vertices receive orthogonal vectors. We prove that for every sufficiently large integer k, it is NP-hard to decide whether the orthogonality dimension of a given graph over ℝ is at most k or at least 2^{(1-o(1))⋅k/2}. We further prove such hardness results for the orthogonality dimension over finite fields as well as for the closely related minrank parameter, which is motivated by the index coding problem in information theory. This in particular implies that it is NP-hard to approximate these graph quantities to within any constant factor. Previously, the hardness of approximation was known to hold either assuming certain variants of the Unique Games Conjecture or for approximation factors smaller than 3/2. The proofs involve the concept of line digraphs and bounds on their orthogonality dimension and on the minrank of their complement.

Abstract. The orthogonality dimension of a graph [Formula: see text] over [Formula: see text] is the smallest integer [Formula: see text] for which one can assign a nonzero [Formula: see text]-dimensional real vector to each vertex of [Formula: see text] such that every two adjacent vertices receive orthogonal vectors. We prove that for every sufficiently large integer [Formula: see text], it is [Formula: see text]-hard to decide whether the orthogonality dimension of a given graph over [Formula: see text] is at most [Formula: see text] or at least [Formula: see text]. We further prove such hardness results for the orthogonality dimension over finite fields as well as for the closely related minrank parameter, which is motivated by the index coding problem in information theory. This in particular implies that it is [Formula: see text]-hard to approximate these graph quantities to within any constant factor. Previously, the hardness of approximation was known to hold either assuming certain variants of the unique games conjecture or for approximation factors smaller than [Formula: see text]. The proofs involve the concept of line digraphs and bounds on their orthogonality dimension and on the minrank of their complement.

The Kneser graph $K(n,k)$ is defined for integers $n$ and $k$ with $n \geq 2k$ as the graph whose vertices are all the $k$-subsets of $\{1,2,\ldots,n\}$ where two such sets are adjacent if they are disjoint. A classical result of Lovász asserts that the chromatic number of $K(n,k)$ is $n-2k+2$. In the computational Kneser problem, we are given an oracle access to a coloring of the vertices of $K(n,k)$ with $n-2k+1$ colors, and the goal is to find a monochromatic edge. We present a randomized algorithm for the Kneser problem with running time $n^{O(1)} \cdot k^{O(k)}$. This shows that the problem is fixed-parameter tractable with respect to the parameter $k$. The analysis involves structural results on intersecting families and on induced subgraphs of Kneser graphs. We also study the Agreeable-Set problem of assigning a small subset of a set of $m$ items to a group of $\ell$ agents, so that all agents value the subset at least as much as its complement. As an application of our algorithm for the Kneser problem, we obtain a randomized polynomial-time algorithm for the Agreeable-Set problem for instances that satisfy $\ell \geq m - O(\frac{\log m}{\log \log m})$. We further show that the Agreeable-Set problem is at least as hard as a variant of the Kneser problem with an extended access to the input coloring.

The Kneser graph $K(n,k)$ is defined for integers $n$ and $k$ with $n \geq 2k$ as the graph whose vertices are all the $k$-subsets of $[n]=\{1,2,\ldots,n\}$ where two such sets are adjacent if they are disjoint. The Schrijver graph $S(n,k)$ is defined as the subgraph of $K(n,k)$ induced by the collection of all $k$-subsets of $[n]$ that do not include two consecutive elements modulo $n$. It is known that the chromatic number of both $K(n,k)$ and $S(n,k)$ is $n-2k+2$. In the computational Kneser and Schrijver problems, we are given an access to a coloring with $n-2k+1$ colors of the vertices of $K(n,k)$ and $S(n,k)$ respectively, and the goal is to find a monochromatic edge. We prove that the problems admit randomized algorithms with running time $n^{O(1)} \cdot k^{O(k)}$, hence they are fixed-parameter tractable with respect to the parameter $k$. The analysis involves structural results on intersecting families and on induced subgraphs of Kneser and Schrijver graphs. We also study the Agreeable-Set problem of assigning a small subset of a set of $m$ items to a group of $\ell$ agents, so that all agents value the subset at least as much as its complement. As an application of our algorithm for the Kneser problem, we obtain a randomized polynomial-time algorithm for the Agreeable-Set problem for instances with $\ell \geq m - O(\frac{\log m}{\log \log m})$. We further show that the Agreeable-Set problem is at least as hard as a variant of the Kneser problem with an extended access to the input coloring.

A $0,1$ matrix is said to be regular if all of its rows and columns have the same number of ones. We prove that for infinitely many integers $k$, there exists a square regular $0,1$ matrix with binary rank $k$, such that the Boolean rank of its complement is $k^{\widetildeΩ(\log k)}$. Equivalently, the ones in the matrix can be partitioned into $k$ combinatorial rectangles, whereas the number of rectangles needed for any cover of its zeros is $k^{\widetildeΩ(\log k)}$. This settles, in a strong form, a question of Pullman (Linear Algebra Appl., 1988) and a conjecture of Hefner, Henson, Lundgren, and Maybee (Congr. Numer., 1990). The result can be viewed as a regular analogue of a recent result of Balodis, Ben-David, Göös, Jain, and Kothari (FOCS, 2021), motivated by the clique vs. independent set problem in communication complexity and by the (disproved) Alon-Saks-Seymour conjecture in graph theory. As an application of the produced regular matrices, we obtain regular counterexamples to the Alon-Saks-Seymour conjecture and prove that for infinitely many integers $k$, there exists a regular graph with biclique partition number $k$ and chromatic number $k^{\widetildeΩ(\log k)}$.

The orthogonality dimension of a graph $G=(V,E)$ over a field $\mathbb{F}$ is the smallest integer $t$ for which there exists an assignment of a vector $u_v \in \mathbb{F}^t$ with $\langle u_v,u_v \rangle \neq 0$ to every vertex $v \in V$, such that $\langle u_v, u_{v'} \rangle = 0$ whenever $v$ and $v'$ are adjacent vertices in $G$. The study of the orthogonality dimension of graphs is motivated by various applications in information theory and in theoretical computer science. The contribution of the present work is two-fold. First, we prove that there exists a constant $c$ such that for every sufficiently large integer $t$, it is $\mathsf{NP}$-hard to decide whether the orthogonality dimension of an input graph over $\mathbb{R}$ is at most $t$ or at least $3t/2-c$. At the heart of the proof lies a geometric result, which might be of independent interest, on a generalization of the orthogonality dimension parameter for the family of Kneser graphs, analogously to a long-standing conjecture of Stahl (J. Comb. Theo. Ser. B, 1976). Second, we study the smallest possible orthogonality dimension over finite fields of the complement of graphs that do not contain certain fixed subgraphs. In particular, we provide an explicit construction of triangle-free $n$-vertex graphs whose complement has orthogonality dimension over the binary field at most $n^{1-δ}$ for some constant $δ>0$. Our results involve constructions from the family of generalized Kneser graphs and they are motivated by the rigidity approach to circuit lower bounds. We use them to answer a couple of questions raised by Codenotti, Pudlák, and Resta (Theor. Comput. Sci., 2000), and in particular, to disprove their Odd Alternating Cycle Conjecture over every finite field.

Let G be a cycle graph and let V₁,…,V_m be a partition of its vertex set into m sets. An independent set S of G is said to fairly represent the partition if |S ∩ V_i| ≥ 1/2⋅|V_i| - 1 for all i ∈ [m]. It is known that for every cycle and every partition of its vertex set, there exists an independent set that fairly represents the partition (Aharoni et al., A Journey through Discrete Math., 2017). We prove that the problem of finding such an independent set is PPA-complete. As an application, we show that the problem of finding a monochromatic edge in a Schrijver graph, given a succinct representation of a coloring that uses fewer colors than its chromatic number, is PPA-complete as well. The work is motivated by the computational aspects of the "cycle plus triangles" problem and of its extensions.

The Boolean rank of a 0,1-matrix A, denoted Rb(A), is the smallest number of monochromatic combinatorial rectangles needed to cover the 1-entries of A. In 1988, de Caen, Gregory, and Pullman asked if the Boolean rank of the Kronecker product Cn ® Cn is strictly smaller than the square of RB(Cn), where Cn is the n x n matrix with zeros on the diagonal and ones everywhere else (Carib. Conf. Comb. & Comp., 1988). A positive answer was given by Watts for n = 4 (Linear Alg. and its Appl., 2001). A result of Karchmer, Kushilevitz, and Nisan, motivated by direct-sum questions in non-deterministic communication complexity, implies that the Boolean rank of Cn ® Cn grows linearly in that of Cn (SIAM J. Disc. Math., 1995), and thus Rb(Cn ® Cn) < RB(Cn)2 for every sufficiently large n. Their proof relies on a probabilistic argument. In this work, we present a general method for proving upper bounds on the Boolean rank of Kronecker products of 0,1-matrices. We use it to affirmatively settle the question of de Caen et al. for all integers n > 7. We further provide an explicit construction of a cover of Cn ® Cn, whose number of rectangles nearly matches the optimal asymptotic bound. Our method for proving upper bounds on the Boolean rank of Kronecker products might find applications in different settings as well. We express its potential applicability by extending it to the wider framework of spanoids, recently introduced by Dvir, Gopi, Gu, and Wigderson (SIAM J. Comput., 2020).

A Maximum Distance Separable code over an alphabet F is defined via an encoding function C : Fk→ Fnthat allows to retrieve a message m ∈ Fkfrom the codeword C(m) even after erasing any n - k of its symbols. The minimum possible alphabet size of general (non-linear) MDS codes for given parameters n and k is unknown and forms one of the central open problems in coding theory. The paper initiates the study of the alphabet size of codes in a generalized setting where the coding scheme is required to handle a pre-specified subset of all possible erasure patterns, naturally represented by an n-vertex k-uniform hypergraph. We relate the minimum possible alphabet size of such codes to the strong chromatic number of the hypergraph and analyze the tightness of the obtained bounds for both the linear and non-linear settings. We further consider variations of the problem which allow a small probability of decoding error.

In the index coding problem a sender holds a message x ∈ {0, 1}nand wishes to broadcast information to n receivers in a way that enables the ith receiver to retrieve the ith bit xi. Every receiver has prior side information comprising a subset of the bits of x, and the goal is to minimize the length of the information sent via the broadcast channel. Porter and Wootters have recently introduced the model of embedded index coding, where the receivers also play the role of the sender and the goal is to minimize the total length of their broadcast information. An embedded index code is said to be task-based if every receiver retrieves its bit based only on the information provided by one of the receivers. This paper studies the effect of the task-based restriction on linear embedded index coding. It is shown that for certain side information maps there exists a linear embedded index code of length quadratically smaller than that of any task-based embedded index code. The result attains, up to a multiplicative constant, the largest possible gap between the two quantities. The proof is by an explicit construction and the analysis involves spectral techniques.

A $t$-dimensional orthogonal representation of a hypergraph is an assignment of nonzero vectors in $\mathbb{R}^t$ to its vertices, such that every hyperedge contains two vertices whose vectors are orthogonal. The orthogonality dimension of a hypergraph $H$, denoted by $\overlineξ(H)$, is the smallest integer $t$ for which there exists a $t$-dimensional orthogonal representation of $H$. In this paper we study computational aspects of the orthogonality dimension of graphs and hypergraphs. We prove that for every $k \geq 4$, it is $\mathsf{NP}$-hard (resp. quasi-$\mathsf{NP}$-hard) to distinguish $n$-vertex $k$-uniform hypergraphs $H$ with $\overlineξ(H) \leq 2$ from those satisfying $\overlineξ(H) \geq Ω(\log^δn)$ for some constant $δ>0$ (resp. $\overlineξ(H) \geq Ω(\log^{1-o(1)} n)$). For graphs, we relate the $\mathsf{NP}$-hardness of approximating the orthogonality dimension to a variant of a long-standing conjecture of Stahl. We also consider the algorithmic problem in which given a graph $G$ with $\overlineξ(G) \leq 3$ the goal is to find an orthogonal representation of $G$ of as low dimension as possible, and provide a polynomial time approximation algorithm based on semidefinite programming.

A set of integers is sum-free if it contains no solution to the equation $x+y=z$. We study sum-free subsets of the set of integers $[n]=\{ 1, \ldots ,n \}$ for which the integer $2n+1$ cannot be represented as a sum of their elements. We prove a bound of $O(2^{n/3})$ on the number of these sets, which matches, up to a multiplicative constant, the lower bound obtained by considering all subsets of $B_n = \{ \lceil \frac{2}{3}(n+1) \rceil, \ldots, n \}$. A main ingredient in the proof is a stability theorem saying that if a subset of $[n]$ of size close to $|B_n|$ contains only a few subsets that contradict the sum-freeness or the forbidden sum, then it is almost contained in $B_n$. Our results are motivated by the question of counting symmetric complete sum-free subsets of cyclic groups of prime order. The proofs involve Freiman's 3k-4 theorem, Green's arithmetic removal lemma, and structural results on independent sets in hypergraphs.

Two classical upper bounds on the Shannon capacity of graphs are the $\vartheta$-function due to Lovász and the minrank parameter due to Haemers. We provide several explicit constructions of $n$-vertex graphs with a constant $\vartheta$-function and minrank at least $n^δ$ for a constant $δ>0$ (over various prime order fields). This implies a limitation on the $\vartheta$-function-based algorithmic approach to approximating the minrank parameter of graphs. The proofs involve linear spaces of multivariate polynomials and the method of higher incidence matrices.

We prove that there exist non-linear binary cyclic codes that attain the Gilbert-Varshamov bound.

A matrix A ∊ ℂq×N satisfies the restricted isometry property of order k with constant ∊ if it preserves the ℓ2 norm of all k-sparse vectors up to a factor of 1 ± ∊. We prove that a matrix A obtained by randomly sampling q = O(k · log2 k · log N) rows from an N × N Fourier matrix satisfies the restricted isometry property of order k with a fixed ∊ with high probability. This improves on Rudelson and Vershynin (Comm. Pure Appl. Math., 2008), its subsequent improvements, and Bourgain (GAFA Seminar Notes, 2014).

A function defined on the Boolean hypercube is $k$-Fourier-sparse if it has at most $k$ nonzero Fourier coefficients. For a function $f: \mathbb{F}_2^n \rightarrow \mathbb{R}$ and parameters $k$ and $d$, we prove a strong upper bound on the number of $k$-Fourier-sparse Boolean functions that disagree with $f$ on at most $d$ inputs. Our bound implies that the number of uniform and independent random samples needed for learning the class of $k$-Fourier-sparse Boolean functions on $n$ variables exactly is at most $O(n \cdot k \log k)$. As an application, we prove an upper bound on the query complexity of testing Booleanity of Fourier-sparse functions. Our bound is tight up to a logarithmic factor and quadratically improves on a result due to Gur and Tamuz (Chicago J. Theor. Comput. Sci., 2013).

A function f : Fn/2 → {0,1} is triangle-free if there are no x1, x2, x3 ∈ Fn/2 satisfying x1 + x2 + x3 --0 and f(x1) -- f(x2) -- f(x3) -- 1. In testing triangle freeness, the goal is to distinguish with high probability triangle-free functions from those which are ε-far from being triangle-free. It was shown by Green that the query complexity of the canonical tester for the problem is upper bounded by a function that depends only on ε (GAFA, 2005), however the best known upper bound is a tower type function of 1/ε. The best known lower bound on the query complexity of the canonical tester is 1/ε13.239 (Fu and Kleinberg, RANDOM, 2014).

We study the Lattice Isomorphism Problem (LIP), in which given two lattices ℒ1 and ℒ2 the goal is to decide whether there exists an orthogonal linear transformation mapping L1 to ℒ2. Our main result is an algorithm for this problem running in time nO(n) times a polynomial in the input size, where n is the rank of the input lattices. A crucial component is a new generalized isolation lemma, which can isolate n linearly independent vectors in a given subset of ℤn and might be useful elsewhere. We also prove that LIP lies in the complexity class SZK.

For a hypergraph H on the vertex set {1,...,n}, a distribution D = (D_1,...,D_n) over {0,1}^n is H-wise independent if every restriction of D to indices which form an edge in H is uniform. This generalizes the notion of k-wise independence obtained by taking H to be the complete n vertex k-uniform hypergraph. This generalization was studied by Schulman (STOC 1992), who presented constructions of H-wise independent distributions that are linear, i.e., the samples are strings of inner products (over F2) of a fixed set of vectors with a uniformly chosen random vector. Let l(H) denote the minimum possible size of a sample space of a uniform H-wise independent distribution. The l parameter is well understood for the special case of k-wise independence. In this work we study the notion of H-wise independence and the l parameter for general graphs and hypergraphs. For graphs, we show how the l parameter relates to standard graph parameters (e.g., clique number, chromatic number, Lovasz theta function, minrank). We derive algorithmic and hardness results for this parameter as well as an explicit construction of graphs G for which l(G) is exponentially smaller than the size of the sample space of any linear G-wise independent distribution. For hypergraphs, we study the problem of testing whether a given distribution is H-wise independent, generalizing results of Alon et al. (STOC 2007).

In the index coding problem, the goal is to transmit an n character word over a field F to n receivers (one character per receiver), where the receivers have side information represented by a graph G. The objective is to minimize the length of a codeword broadcasted to all receivers which allows each receiver to learn its character. For linear index coding, the minimum possible length is known to be equal to the minrank parameter. In this paper we initiate the study of the typical minimum length of a linear index code for the random graph G(n, p) over a field F. First, we prove that for every constant size field F and a constant p, the minimum length of a linear index code for G(n, p) over F is almost surely Ω(√n). Second, we introduce and study two special models of index coding and study their typical minimum length: Locally decodable index codes in which the receivers are required to query at most q characters from the encoded message (such codes naturally correspond to efficient decoding); and low density index codes in which every character of the broadcasted word affects at most q characters in the encoded message (such codes naturally correspond to efficient encoding procedures). We present enhanced results for these special models.

In the index coding problem, introduced by Birk and Kol (INFOCOM, 1998), the goal is to broadcast an n bit word to n receivers (one bit per receiver), where the receivers have side information represented by a graph G. The objective is to minimize the length of a codeword sent to all receivers which allows each receiver to learn its bit. For linear index coding, the minimum possible length is known to be equal to a graph parameter called minrank (Bar-Yossef et al., FOCS, 2006). We show a polynomial time algorithm that, given an n vertex graph G with minrank k, finds a linear index code for G of length Õ(nf(k))), where f(k) depends only on k. For example, for k = 3 we obtain f(3) ≈ 0.2574. Our algorithm employs a semidefinite program (SDP) introduced by Karger, Motwani and Sudan (J. ACM, 1998) for graph coloring and its refined analysis due to Arora, Chlamtac and Charikar (STOC, 2006). Since the SDP we use is not a relaxation of the minimization problem we consider, a crucial component of our analysis is an upper bound on the objective value of the SDP in terms of the minrank. At the heart of our analysis lies a combinatorial result which may be of independent interest. Namely, we show an exact expression for the maximum possible value of the Lovász ϑ-function of a graph with minrank k. This yields a tight gap between two classical upper bounds on the Shannon capacity of a graph.

In the online channel coding model, a sender wishes to communicate a message to a receiver by transmitting a codeword x =(x_1,...,x_n) in {0,1}^n bit by bit via a channel limited to at most pn corruptions. The channel is online in the sense that at the ith step the channel decides whether to flip the ith bit or not and its decision is based only on the bits transmitted so far, i.e., (x_1,...,x_i). This is in contrast to the classical adversarial channel in which the corruption is chosen by a channel that has full knowledge on the sent codeword x. The best known lower bound on the capacity of both the online channel and the classical adversarial channel is the well-known Gilbert-Varshamov bound. In this paper we prove a lower bound on the capacity of the online channel which beats the Gilbert-Varshamov bound for any positive p such that H(2p) < 0.5 (where H is the binary entropy function). To do so, we prove that for any such p, a code chosen at random combined with the nearest neighbor decoder achieves with high probability a rate strictly higher than the Gilbert-Varshamov bound (for the online channel).

We show a connection between the semidefinite relaxation of unique games and their behavior under parallel repetition. Specifically,denoting by val(G) the value of a two-prover unique game G, andby sdpval(G) the value of a natural semidefinite program to approximate val(G), we prove that for every l epsi N, if sdpval(G) ges 1-delta, then val(Gl) ges 1-radicsldelta. Here, Gldenotes the l-fold parallel repetition of G, and s=O(log(k/delta)), where k denotes the alphabet size of the game. For the special case where G is an XOR game (i.e., k=2), we obtain the same bound but with s as an absolute constant. Our bounds on s are optimal up to a factor of O(log(1/delta)). For games with a significant gap between the quantities val(G) and sdpval(G), our result implies that val(Gl) may be much larger than val(G)l, giving a counterexample to the strong parallel repetition conjecture. In a recent breakthrough, Raz (FOCS'08) has shown such an example using the max-cut game on oddcycles. Our results are based on a generalization of his techniques.

We show that unless NP ⊆ RTIME (2poly(log n)), for any ε > 0 there is no polynomial-time algorithm approximating the Shortest Vector Problem (SVP) on n-dimensional lattices inthe lp norm (1 ≤q p<∞) to within a factor of 2(log n)1-ε. This improves the previous best factor of 2(logn)1/2-ε under the same complexity assumption due to Khot. Under the stronger assumption NP ࣰ RSUBEXP, we obtain a hardness factor of nc/log log n for some c > 0.

We provide the first hardness result for the covering radius problem on lattices (CRP). Namely, we show that for any large enough p les infin there exists a constant cp> 1 such that CRP in the lscrpnorm is Pi2-hard to approximate to within any constant less than cp. In particular, for the case p = infin, we obtain the constant Cinfin= 1.5. This gets close to the constant 2 beyond which the problem is not believed to be Pi2-hard. As part of our proof, we establish a stronger hardness of approximation result for the forallexist-3-SAT problem with bounded occurrences. This hardness result might be useful elsewhere