On the Range of the Permanent of $(\pm1)$-Matrices
We establish a superpolynomial lower bound on the range of the permanent function on the set of $n\times n$ matrices with $\pm1$ entries.
math.CO↗
arXiv subjects
Alexander Razborov
Publications and source records attributed to Alexander Razborov.
On Domination Exponents for Pairs of Graphs
Understanding graph density profiles is notoriously challenging. Even for pairs of graphs, complete characterizations are known only in very limited cases, such as edges versus cliques. This paper explores a relaxation of the graph density profile problem by examining the homomorphism density domination exponent $C(H_1, H_2)$. This is the smallest real number $c \geq 0$ such that $t(H_1, T) \geq t(H_2, T)^c$ for all target graphs $T$ (if such a $c$ exists) where $t(H,T)$ is the homomorphism density from $H$ to $T$. We demonstrate that infinitely many families of graphs are required to realize $C(H_1, H_2)$ for all connected graphs $H_1$, $H_2$. We derive the homomorphism density domination exponent for a variety of graph pairs, including paths and cycles. As a couple of typical examples, we obtain exact values when $H_1$ is an even cycle and $H_2$ contains a Hamiltonian cycle, and provide asymptotically sharp bounds when both $H_1$ and $H_2$ are odd cycles.
math.CO↗
Improved Convergence Guarantees for Shallow Neural Networks
We continue a long line of research aimed at proving convergence of depth 2 neural networks, trained via gradient descent, to a global minimum. Like in many previous works, our model has the following features: regression with quadratic loss function, fully connected feedforward architecture, RelU activations, Gaussian data instances and network initialization, adversarial labels. It is more general in the sense that we allow both layers to be trained simultaneously and at {\em different} rates. Our results improve on state-of-the-art [Oymak Soltanolkotabi 20] (training the first layer only) and [Nguyen 21, Section 3.2] (training both layers with Le Cun's initialization). We also report several simple experiments with synthetic data. They strongly suggest that, at least in our model, the convergence phenomenon extends well beyond the ``NTK regime''.
cs.LG↗
More about sparse halves in triangle-free graphs
One of Erdos's conjectures states that every triangle-free graph on $n$ vertices has an induced subgraph on $n/2$ vertices with at most $n^2/50$ edges. We report several partial results towards this conjecture. In particular, we establish the new bound $\frac{27}{1024}n^2$ on the number of edges in general case. We completely prove the conjecture for graphs of girth $\geq 5$, for graphs with independence number $\geq 2n/5$ and for strongly regular graphs. Each of these three classes includes both known (conjectured) extremal configurations, the 5-cycle and the Petersen graph.
math.CO↗
Clique Is Hard on Average for Regular Resolution
We prove that for $k \ll \sqrt[4]{n}$ regular resolution requires length $n^{Ω(k)}$ to establish that an Erdős-Rényi graph with appropriately chosen edge density does not contain a $k$-clique. This lower bound is optimal up to the multiplicative constant in the exponent, and also implies unconditional $n^{Ω(k)}$ lower bounds on running time for several state-of-the-art algorithms for finding maximum cliques in graphs.
cs.CC↗
An Extremal Problem Motivated by Triangle-Free Strongly Regular Graphs
We introduce the following combinatorial problem. Let $G$ be a triangle-free regular graph with edge density $ρ$. What is the minimum value $a(ρ)$ for which there always exist two non-adjacent vertices such that the density of their common neighborhood is $\leq a(ρ)$? We prove a variety of upper bounds on the function $a(ρ)$ that are tight for the values $ρ=2/5,\ 5/16,\ 3/10,\ 11/50$, with $C_5$, Clebsch, Petersen and Higman-Sims being respective extremal configurations. Our proofs are entirely combinatorial and are largely based on counting densities in the style of flag algebras. For small values of $ρ$, our bound attaches a combinatorial meaning to Krein conditions that might be interesting in its own right. We also prove that for any $ε>0$ there are only finitely many values of $ρ$ with $a(ρ)\geqε$ but this finiteness result is somewhat purely existential (the bound is double exponential in $1/ε$).
math.CO↗
Polynomial to exponential transition in Ramsey theory
Given $s \ge k\ge 3$, let $h^{(k)}(s)$ be the minimum $t$ such that there exist arbitrarily large $k$-uniform hypergraphs $H$ whose independence number is at most polylogarithmic in the number of vertices and in which every $s$ vertices span at most $t$ edges. Erd\H os and Hajnal conjectured (1972) that $h^{(k)}(s)$ can be calculated precisely using a recursive formula and Erd\H os offered \$500 for a proof of this. For $k=3$ this has been settled for many values of $s$ including powers of three but it was not known for any $k\geq 4$ and $s\geq k+2$. Here we settle the conjecture for all $s \ge k \ge 4$. We also answer a question of Bhat and Rödl by constructing, for each $k \ge 4$, a quasirandom sequence of $k$-uniform hypergraphs with positive density and upper density at most $k!/(k^k-k)$. This result is sharp.
math.CO↗
On CDCL-based proof systems with the ordered decision strategy
We prove that conflict-driven clause learning SAT-solvers with the ordered decision strategy and the DECISION learning scheme are equivalent to ordered resolution. We also prove that, by replacing this learning scheme with its opposite that stops after the first new clause when backtracking, they become equivalent to general resolution. To the best of our knowledge, this is the first theoretical study of the interplay between specific decision strategies and clause learning. For both results, we allow nondeterminism in the solver's ability to perform unit propagation, conflict analysis, and restarts, in a way that is similar to previous works in the literature. To aid the presentation of our results, and possibly future research, we define a model and language for discussing CDCL-based proof systems that allows for succinct and precise theorem statements.
cs.LO↗
Asymptotic Structure of Graphs with the Minimum Number of Triangles
We consider the problem of minimizing the number of triangles in a graph of given order and size and describe the asymptotic structure of extremal graphs. This is achieved by characterizing the set of flag algebra homomorphisms that minimize the triangle density.
math.CO↗
On the Number of Pentagons in Triangle-Free Graphs
Using the formalism of flag algebras, we prove that every triangle-free graph $G$ with $n$ vertices contains at most $(n/5)^5$ cycles of length five. Moreover, the equality is attained only when $n$ is divisible by five and $G$ is the balanced blow-up of the pentagon. We also compute the maximal number of pentagons and characterize extremal graphs in the non-divisible case provided $n$ is sufficiently large. This settles a conjecture made by Erdős in 1984.
math.CO↗
On Turan's (3,4)-problem with forbidden configurations
We identify three 3-graphs on five vertices each missing in all known extremal configurations for Turan's (3,4)-problem and prove Turan's conjecture for 3-graphs that are additionally known not to contain any induced copies of these 3-graphs. Our argument is based on an (apparently) new technique of "indirect interpretation" that allows us to retrieve additional structure from hypothetical counterexamples to Turan's conjecture, but in rather loose and limited sense. We also include two miscellaneous calculations in flag algebras that prove similar results about some other additional forbidden subgraphs.
math.CO↗
On the Caccetta-Haggkvist conjecture with forbidden subgraphs
The Caccetta-Haggkvist conjecture made in 1978 asserts that every orgraph on n vertices without oriented cycles of length <= l must contain a vertex of outdegree at most (n-1)/l. It has a rather elaborate set of (conjectured) extremal configurations. In this paper we consider the case l=3 that received quite a significant attention in the literature. We identify three orgraphs on four vertices each that are missing as an induced subgraph in all known extremal examples and prove the Caccetta-Haggkvist conjecture for orgraphs missing as induced subgraphs any of these orgraphs, along with cycles of length 3. Using a standard trick, we can also lift the restriction of being induced, although this makes graphs in our list slightly more complicated.
math.CO↗
Non-three-colorable common graphs exist
A graph H is called common if the total number of copies of H in every graph and its complement asymptotically minimizes for random graphs. A former conjecture of Burr and Rosta, extending a conjecture of Erdos asserted that every graph is common. Thomason disproved both conjectures by showing that the complete graph of order four is not common. It is now known that in fact the common graphs are very rare. Answering a question of Sidorenko and of Jagger, Stovicek and Thomason from 1996 we show that the 5-wheel is common. This provides the first example of a common graph that is not three-colorable.
math.CO↗
On the Fon-der-Flaass Interpretation of Extremal Examples for Turan's (3,4)-problem
In 1941, Turan conjectured that the edge density of any 3-graph without independent sets on 4 vertices (Turan (3,4)-graph) is >= 4/9(1-o(1)), and he gave the first example witnessing this bound. Brown (1983) and Kostochka (1982) found many other examples of this density. Fon-der-Flaass (1988) presented a general construction that converts an arbitrary $\vec C_4$-free orgraph $Γ$ into a Turan (3,4)-graph. He observed that all Turan-Brown-Kostochka examples result from his construction, and proved the bound >= 3/7(1-o(1)) on the edge density of any Turan (3,4)-graph obtainable in this way. In this paper we establish the optimal bound 4/9(1-o(1)) on the edge density of any Turan (3,4)-graph resulting from the Fon-der-Flaass construction under any of the following assumptions on the undirected graph $G$ underlying the orgraph $Γ$: 1. $G$ is complete multipartite; 2. The edge density of $G$ is >= (2/3-epsilon) for some absolute constant epsilon>0. We are also able to improve Fon-der-Flaass's bound to 7/16(1-o(1)) without any extra assumptions on $Γ$.
math.CO↗
On Minimal Unsatisfiability and Time-Space Trade-offs for k-DNF Resolution
In the context of proving lower bounds on proof space in k-DNF resolution, [Ben-Sasson and Nordstrom 2009] introduced the concept of minimally unsatisfiable sets of k-DNF formulas and proved that a minimally unsatisfiable k-DNF set with m formulas can have at most O((mk)^(k+1)) variables. They also gave an example of such sets with Omega(mk^2) variables. In this paper we significantly improve the lower bound to Omega(m)^k, which almost matches the upper bound above. Furthermore, we show that this implies that the analysis of their technique for proving time-space separations and trade-offs for k-DNF resolution is almost tight. This means that although it is possible, or even plausible, that stronger results than in [Ben-Sasson and Nordstrom 2009] should hold, a fundamentally different approach would be needed to obtain such results.
cs.DM↗
Almost Euclidean subspaces of \ell_1^N via expander codes
We give an explicit (in particular, deterministic polynomial time) construction of subspaces X of R^N of dimension (1-o(1))N such that for every element x in X, |x|_1 and N^{1/2} |x|_2 are equivalent up to a factor of (log N)^{log log log N}. If we are allowed to use N^{o(1)} random bits, this factor can be improved to poly(log N). Our construction makes use of unbalanced bipartite graphs to impose local linear constraints on vectors in the subspace, and our analysis relies on expansion properties of the graph. This is inspired by similar constructions of error-correcting codes.
math.MG↗
Quantum communication complexity of symmetric predicates
We completely (that is, up to a logarithmic factor) characterize the bounded-error quantum communication complexity of every predicate $f(x,y)$ depending only on $|x\cap y|$ ($x,y\subseteq [n]$). Namely, for a predicate $D$ on $\{0,1,...,n\}$ let $\ell_0(D)\df \max\{\ell : 1\leq\ell\leq n/2\land D(\ell)\not\equiv D(\ell-1)\}$ and $\ell_1(D)\df \max\{n-\ell : n/2\leq\ell < n\land D(\ell)\not\equiv D(\ell+1)\}$. Then the bounded-error quantum communication complexity of $f_D(x,y) = D(|x\cap y|)$ is equal (again, up to a logarithmic factor) to $\sqrt{n\ell_0(D)}+\ell_1(D)$. In particular, the complexity of the set disjointness predicate is $Ω(\sqrt n)$. This result holds both in the model with prior entanglement and without it.
quant-ph↗