arXiv subjects

Marcos Kiwi

Publications and source records attributed to Marcos Kiwi.

At least 19 recordsLinked to original sources

arXiv2026

About a ball removal process on bins

We consider a basic balls-and-bins question in which you have to distribute $n$ balls into $k$ bins. Then, round by round, a ball is removed from a non-empty bin chosen uniformly at random. The process ends when a single non-empty bin remains. The goal is to minimize the expected number of remaining balls. An open problem posed by Will Ma asks whether the initial assignment that minimizes the expected number of remaining balls is one that is as balanced as possible. Using a coupling argument, we answer this conjecture positively, and we discuss the case of non-uniform choice among the non-empty bins.

math.PR↗

arXiv2025

Mixing time and isoperimetry in random geometric graphs

In this paper we study the mixing time of the simple random walk on the giant component of supercritical $d$-dimensional random geometric graphs generated by the unit intensity Poisson Point Process in a $d$-dimensional cube of volume $n$. With $r_g$ denoting the threshold for having a giant component, we show that for every $\epsilon > 0$ and any $r \ge (1+\epsilon)r_g$, the mixing time of the giant component is with high probability $\Theta(n^{2/d}/r^{2})$, thereby closing a gap in the literature. The main tool is an isoperimetric inequality which holds, w.h.p., for any large enough vertex set, a result which we believe is of independent interest. Our analysis also implies that the relaxation time is of the same order.

math.PR↗

arXiv2024

Naively Sorting Evolving Data is Optimal and Robust

We study sorting in the evolving data model, introduced by [AKMU11], where the true total order changes while the sorting algorithm is processing the input. More precisely, each comparison operation of the algorithm is followed by a sequence of evolution steps, where an evolution step perturbs the rank of a random item by a "small" random value. The goal is to maintain an ordering that remains close to the true order over time. Previous works have analyzed adaptations of classic sorting algorithms, assuming that an evolution step changes the rank of an item by just one, and that a fixed constant number $b$ of evolution steps take place between two comparisons. In fact, the only previous result achieving optimal linear total deviation, by [BvDEGJ18a], applies just for $b=1$. We analyze a very simple sorting algorithm suggested by [M14], which samples a random pair of adjacent items in each step and swaps them if they are out of order. We show that the algorithm achieves and maintains, with high probability, optimal total deviation, $O(n)$, and optimal maximum deviation, $O(\log n)$, under very general model settings. Namely, the perturbation introduced by each evolution step is sampled from a general distribution of bounded moment generating function, and we just require that the average number of evolution steps between two sorting steps be bounded by an (arbitrary) constant, where the average is over a linear number of steps. The key ingredients of our proof are a novel potential function argument that inserts "gaps" in the list of items, and a general analysis framework which separates the analysis of sorting from that of the evolution steps, and is applicable to a variety of settings for which previous approaches do not apply. Our results settle conjectures and open problems in the aforementioned works, and provide theoretical support for empirical observations in [BvDEGJ18b].

cs.DS↗

arXiv2023

Label propagation on binomial random graphs

We study the behavior of a label propagation algorithm (LPA) on the Erd\H{o}s-R\'enyi random graph $\mathcal{G}(n,p)$. Initially, given a network, each vertex starts with a random label in the interval $[0,1]$. Then, in each round of LPA, every vertex switches its label to the majority label in its neighborhood (including its own label). At the first round, ties are broken towards smaller labels, while at each of the next rounds, ties are broken uniformly at random. The algorithm terminates once all labels stay the same in two consecutive iterations. LPA is successfully used in practice for detecting communities in networks (corresponding to vertex sets with the same label after termination of the algorithm). Perhaps surprisingly, LPA's performance on dense random graphs is hard to analyze, and so far convergence to consensus was known only when $np\ge n^{3/4+\varepsilon}$, where LPA converges in three rounds. By defining an alternative label attribution procedure which converges to the label propagation algorithm after three rounds, a careful multi-stage exposure of the edges allows us to break the $n^{3/4+\varepsilon}$ barrier and show that, when $np \ge n^{5/8+\varepsilon}$, a.a.s.\ the algorithm terminates with a single label. Moreover, we show that, if $np\gg n^{2/3}$, a.a.s.\ this label is the smallest one, whereas if $n^{5/8+\varepsilon}\le np\ll n^{2/3}$, the surviving label is a.a.s.\ not the smallest one. En passant, we show a presumably new monotonicity lemma for Binomial random variables that might be of independent interest.

math.PR↗

arXiv2022

Cover and Hitting Times of Hyperbolic Random Graphs

We study random walks on the giant component of Hyperbolic Random Graphs (HRGs), in the regime when the degree distribution obeys a power law with exponent in the range $(2,3)$. In particular, we first focus on the expected time for a random walk to hit a given vertex or visit, i.e. cover, all vertices. We show that, a.a.s. (with respect to the HRG), and up to multiplicative constants: the cover time is $n(\log n)^2$, the maximum hitting time is $n\log n$, and the average hitting time is $n$. We then determine the expected time to commute between two given vertices a.a.s., up to a small factor polylogarithmic in $n$, and under some mild hypothesis on the pair of vertices involved. Our results are proved by controlling effective resistances using the energy dissipated by carefully designed network flows associated to a tiling of the hyperbolic plane, on which we overlay a forest-like structure.

math.PR↗

arXiv2022

Tail bounds for detection times in mobile hyperbolic graphs

Motivated by Krioukov et al.'s model of random hyperbolic graphs for real-world networks, and inspired by the analysis of a dynamic model of graphs in Euclidean space by Peres et al., we introduce a dynamic model of hyperbolic graphs in which vertices are allowed to move according to a Brownian motion maintaining the distribution of vertices in hyperbolic space invariant. For different parameters of the speed of angular and radial motion, we analyze tail bounds for detection times of a fixed target and obtain a complete picture, for very different regimes, of how and when the target is detected: as a function of the time passed, we characterize the subset of the hyperbolic space where particles typically detecting the target are initially located. We overcome several substantial technical difficulties not present in Euclidean space, and provide a complete picture on tail bounds. On the way, we obtain also new results for the time more general continuous processes with drift and reflecting barrier spent in certain regions, and we also obtain improved bounds for independent sums of Pareto random variables.

math.PR↗

arXiv2021

Quasi-random words and limits of word sequences

Words are sequences of letters over a finite alphabet. We study two intimately related topics for this object: quasi-randomness and limit theory. With respect to the first topic we investigate the notion of uniform distribution of letters over intervals, and in the spirit of the famous Chung--Graham--Wilson theorem for graphs we provide a list of word properties which are equivalent to uniformity. In particular, we show that uniformity is equivalent to counting 3-letter subsequences. Inspired by graph limit theory we then investigate limits of convergent word sequences, those in which all subsequence densities converge. We show that convergent word sequences have a natural limit, namely Lebesgue measurable functions of the form $f:[0,1]\to[0,1]$. Via this theory we show that every hereditary word property is testable, address the problem of finite forcibility for word limits and establish as a byproduct a new model of random word sequences. Along the lines of the proof of the existence of word limits, we can also establish the existence of limits for higher dimensional structures. In particular, we obtain an alternative proof of the result by Hoppen, Kohayakawa, Moreira, Ráth and Sampaio [{\it J. Combin. Theory Ser. B 103(1):93--113, 2013}] establishing the existence of permutons.

math.CO↗

arXiv2019

On the second largest component of random hyperbolic graphs

We show that in the random hyperbolic graph model as formalized by Gugelmann et al. in the most interesting range of $\frac12 < α< 1$ the size of the second largest component is $Θ((\log n)^{1/(1-α)})$, thus answering a question of Bode et al. We also show that for $α=\frac12$ with constant probability the corresponding size is $Θ(\log n)$, whereas for $α=1$ it is $Ω(n^{b})$ for some $b > 0$.

math.PR↗

arXiv2017

Spectral Gap of Random Hyperbolic Graphs and Related Parameters

Random hyperbolic graphs have been suggested as a promising model of social networks. A few of their fundamental parameters have been studied. However, none of them concerns their spectra. We consider the random hyperbolic graph model as formalized by [GPP12] and essentially determine the spectral gap of their normalized Laplacian. Specifically, we establish that with high probability the second smallest eigenvalue of the normalized Laplacian of the giant component of and $n$-vertex random hyperbolic graph is $Ω(n^{-(2α-1)}/D)$, where $\frac12<α<1$ is a model parameter and $D$ is the network diameter (which is known to be at most polylogarithmic in $n$). We also show a matching (up to a polylogarithmic factor) upper bound of $n^{-(2α-1)}(\log n)^{1+o(1)}$. As a byproduct we conclude that the conductance upper bound on the eigenvalue gap obtained via Cheeger's inequality is essentially tight. We also provide a more detailed picture of the collection of vertices on which the bound on the conductance is attained, in particular showing that for all subsets whose volume is $O(n^{1-\varepsilon})$ the obtained conductance is with high probability $Ω(n^{-(2α-1)\varepsilon+o(1)})$. Finally, we also show consequences of our result for the minimum and maximum bisection of the giant component.

math.PR↗

arXiv2016

Types of perfect matchings in toroidal square grids

Let $T_{m,n}$ be toroidal square grid of size $m\times n$ and let both $m$ and $n$ be even. Let $P$ be a perfect matching of $T_{m,n}$ and let $D(P)$ be the cycle-rooted spanning forest of $P$ obtained by the generalized Temperley's construction. The types of $P$ and $D(P)$ in the first homology group $H_1(\mathbb{T},\mathbb{Z})$ of torus $\mathbb{T}$ with coefficients in $\mathbb{Z}$ has been extensively studied. In this paper we study the types of $P$ and $D(P)$ in the first homology group $H_1(\mathbb{T},\mathbb{F}_2)$ with the coefficients in $\mathbb{F}_2$. Our considerations connect two remarkable results concerning perfect matchings of toroidal square grids, namely Temperley's bijection and the Arf-invariant formula.

math-ph↗

arXiv2014

A bound for the diameter of random hyperbolic graphs

Random hyperbolic graphs were recently introduced by Krioukov et. al. [KPKVB10] as a model for large networks. Gugelmann, Panagiotou, and Peter [GPP12] then initiated the rigorous study of random hyperbolic graphs using the following model: for $α> \tfrac{1}{2}$, $C\in\mathbb{R}$, $n\in\mathbb{N}$, set $R=2\ln n+C$ and build the graph $G=(V,E)$ with $|V|=n$ as follows: For each $v\in V$, generate i.i.d. polar coordinates $(r_{v},θ_{v})$ using the joint density function $f(r,θ)$, with $θ_{v}$ chosen uniformly from $[0,2π)$ and $r_{v}$ with density $f(r)=\frac{α\sinh(αr)}{\cosh(αR)-1}$ for $0\leq r< R$. Then, join two vertices by an edge, if their hyperbolic distance is at most $R$. We prove that in the range $\tfrac{1}{2} < α< 1$ a.a.s. for any two vertices of the same component, their graph distance is $O(\log^{C_0+1+o(1)}n)$, where $C_0=2/(\tfrac{1}{2}-\frac{3}{4}α+\tfrac{α^2}{4})$, thus answering a question raised in [GPP12] concerning the diameter of such random graphs. As a corollary from our proof we obtain that the second largest component has size $O(\log^{2C_0+1+o(1)}n)$, thus answering a question of Bode, Fountoulakis and Müller [BFM13]. We also show that a.a.s. there exist isolated components forming a path of length $Ω(\log n)$, thus yielding a lower bound on the size of the second largest component.

math.CO↗

arXiv2013

Strict majority bootstrap percolation in the r-wheel

In this paper we study the strict majority bootstrap percolation process on graphs. Vertices may be active or passive. Initially, active vertices are chosen independently with probability p. Each passive vertex becomes active if at least half of its neighbors are active (and thereafter never changes its state). If at the end of the process all vertices become active then we say that the initial set of active vertices percolates on the graph. We address the problem of finding graphs for which percolation is likely to occur for small values of p. Specifically, we study a graph that we call r-wheel: a ring of n vertices augmented with a universal vertex where each vertex in the ring is connected to its r closest neighbors to the left and to its r closest neighbors to the right. We prove that the critical probability is 1/4. In other words, if p>1/4 then for large values of r percolation occurs with probability arbitrarily close to 1 as n goes to infinity. On the other hand, if p<1/4 then the probability of percolation is bounded away from 1.

cs.SI↗

arXiv2013

Repetition-free longest common subsequence of random sequences

A repetition free Longest Common Subsequence (LCS) of two sequences x and y is an LCS of x and y where each symbol may appear at most once. Let R denote the length of a repetition free LCS of two sequences of n symbols each one chosen randomly, uniformly, and independently over a k-ary alphabet. We study the asymptotic, in n and k, behavior of R and establish that there are three distinct regimes, depending on the relative speed of growth of n and k. For each regime we establish the limiting behavior of R. In fact, we do more, since we actually establish tail bounds for large deviations of R from its limiting behavior. Our study is motivated by the so called exemplar model proposed by Sankoff (1999) and the related similarity measure introduced by Adi et al. (2007). A natural question that arises in this context, which as we show is related to long standing open problems in the area of probabilistic combinatorics, is to understand the asymptotic, in n and k, behavior of parameter R.

math.CO↗

arXiv2011

Computational Hardness of Enumerating Satisfying Spin-Assignments in Triangulations

Satisfying spin-assignments in triangulations of a surface are states of minimum energy of the antiferromagnetic Ising model on triangulations which correspond (via geometric duality) to perfect matchings in cubic bridgeless graphs. In this work we show that it is NP-complete to decide whether or not a surface triangulation admits a satisfying spin-assignment, and that it is #P-complete to determine the number of such assignments. Both results are derived via an elaborate (and atypical) reduction that maps a Boolean formula in 3-conjunctive normal form into a triangulation of an orientable closed surface.

cs.CC↗

arXiv2011

Generalizations and Variants of the Largest Non-crossing Matching Problem in Random Bipartite Graphs

We are interested in the statistics of the length of the longest increasing subsequence of 2-rowed lexicographically sorted arrays chosen according to distinct families of distributions D = (D_n)_n, and when n goes to infinity. This framework encompasses well studied problems such as the so called Longest Increasing Subsequence problem, the Longest Common Subsequence problem, problems concerning directed bond percolation models, among others. We define several natural families of distinct distributions and characterize the asymptotic behavior of the expected length of a longest increasing subsequence chosen according to them. In particular, we consider generalizations to d-rowed arrays as well as symmetry restricted two-rowed arrays.

math.CO↗

arXiv2010

Counting perfect matchings of cubic graphs in the geometric dual

Lovász and Plummer conjectured, in the mid 1970's, that every cubic graph G with no cutedge has an exponential in |V(G)| number of perfect matchings. In this work we show that every cubic planar graph G whose geometric dual graph is a stack triangulation has at least 3 times the golden ratio to |V(G)|/72 distinct perfect matchings. Our work builds on a novel approach relating Lovász and Plummer's conjecture and the number of so called groundstates of the widely studied Ising model from statistical physics.

math.CO↗

arXiv2009

Satisfying states of triangulations of a convex n-gon

In this work we count the number of satisfying states of triangulations of a convex n-gon using the transfer matrix method. We show an exponential (in n) lower bound. We also give the exact formula for the number of satisfying states of a strip of triangles.

math.CO↗

arXiv2008

On a Speculated Relation Between Chvátal-Sankoff Constants of Several Sequences

It is well known that, when normalized by n, the expected length of a longest common subsequence of d sequences of length n over an alphabet of size sigma converges to a constant gamma_{sigma,d}. We disprove a speculation by Steele regarding a possible relation between gamma_{2,d} and gamma_{2,2}. In order to do that we also obtain new lower bounds for gamma_{sigma,d}, when both sigma and d are small integers.

math.CO↗