arXiv subjects

Lennart Gulikers

Publications and source records attributed to Lennart Gulikers.

6 recordsLinked to original sources

arXiv2018

An Impossibility Result for Reconstruction in a Degree-Corrected Planted-Partition Model

We consider the Degree-Corrected Stochastic Block Model (DC-SBM): a random graph on $n$ nodes, having i.i.d. weights $(ϕ_u)_{u=1}^n$ (possibly heavy-tailed), partitioned into $q \geq 2$ asymptotically equal-sized clusters. The model parameters are two constants $a,b > 0$ and the finite second moment of the weights $Φ^{(2)}$. Vertices $u$ and $v$ are connected by an edge with probability $\frac{ϕ_u ϕ_v}{n}a$ when they are in the same class and with probability $\frac{ϕ_u ϕ_v}{n}b$ otherwise. We prove that it is information-theoretically impossible to estimate the clusters in a way positively correlated with the true community structure when $(a-b)^2 Φ^{(2)} \leq q(a+b)$. As by-products of our proof we obtain $(1)$ a precise coupling result for local neighbourhoods in DC-SBM's, that we use in a follow up paper [Gulikers et al., 2017] to establish a law of large numbers for local-functionals and $(2)$ that long-range interactions are weak in (power-law) DC-SBM's.

math.PR↗

arXiv2018

Adaptive Matching for Expert Systems with Uncertain Task Types

A matching in a two-sided market often incurs an externality: a matched resource may become unavailable to the other side of the market, at least for a while. This is especially an issue in online platforms involving human experts as the expert resources are often scarce. The efficient utilization of experts in these platforms is made challenging by the fact that the information available about the parties involved is usually limited. To address this challenge, we develop a model of a task-expert matching system where a task is matched to an expert using not only the prior information about the task but also the feedback obtained from the past matches. In our model the tasks arrive online while the experts are fixed and constrained by a finite service capacity. For this model, we characterize the maximum task resolution throughput a platform can achieve. We show that the natural greedy approaches where each expert is assigned a task most suitable to her skill is suboptimal, as it does not internalize the above externality. We develop a throughput optimal backpressure algorithm which does so by accounting for the `congestion' among different task types. Finally, we validate our model and confirm our theoretical findings with data-driven simulations via logs of Math.StackExchange, a StackOverflow forum dedicated to mathematics.

cs.AI↗

arXiv2017

Non-Backtracking Spectrum of Degree-Corrected Stochastic Block Models

Motivated by community detection, we characterise the spectrum of the non-backtracking matrix $B$ in the Degree-Corrected Stochastic Block Model. Specifically, we consider a random graph on $n$ vertices partitioned into two equal-sized clusters. The vertices have i.i.d. weights $\{ ϕ_u \}_{u=1}^n$ with second moment $Φ^{(2)}$. The intra-cluster connection probability for vertices $u$ and $v$ is $\frac{ϕ_u ϕ_v}{n}a$ and the inter-cluster connection probability is $\frac{ϕ_u ϕ_v}{n}b$. We show that with high probability, the following holds: The leading eigenvalue of the non-backtracking matrix $B$ is asymptotic to $ρ= \frac{a+b}{2} Φ^{(2)}$. The second eigenvalue is asymptotic to $μ_2 = \frac{a-b}{2} Φ^{(2)}$ when $μ_2^2 > ρ$, but asymptotically bounded by $\sqrtρ$ when $μ_2^2 \leq ρ$. All the remaining eigenvalues are asymptotically bounded by $\sqrtρ$. As a result, a clustering positively-correlated with the true communities can be obtained based on the second eigenvector of $B$ in the regime where $μ_2^2 > ρ.$ In a previous work we obtained that detection is impossible when $μ_2^2 < ρ,$ meaning that there occurs a phase-transition in the sparse regime of the Degree-Corrected Stochastic Block Model. As a corollary, we obtain that Degree-Corrected Erdős-Rényi graphs asymptotically satisfy the graph Riemann hypothesis, a quasi-Ramanujan property. A by-product of our proof is a weak law of large numbers for local-functionals on Degree-Corrected Stochastic Block Models, which could be of independent interest.

math.PR↗

arXiv2017

A spectral method for community detection in moderately-sparse degree-corrected stochastic block models

We consider community detection in Degree-Corrected Stochastic Block Models (DC-SBM). We propose a spectral clustering algorithm based on a suitably normalized adjacency matrix. We show that this algorithm consistently recovers the block-membership of all but a vanishing fraction of nodes, in the regime where the lowest degree is of order log$(n)$ or higher. Recovery succeeds even for very heterogeneous degree-distributions. The used algorithm does not rely on parameters as input. In particular, it does not need to know the number of communities.

math.PR↗

arXiv2015

Explosiveness of Age-Dependent Branching Processes with Contagious and Incubation Periods

We study explosiveness of age-dependent branching processes describing the early stages of an epidemic-spread: both forward- and backward process are analysed. For the classical age-dependent branching process $(h,G)$, where the offspring has probability generating function $h$ and all individuals have life-lengths independently picked from a distribution $G$, we focus on the setting $h = h_α^L$, with $L$ a function varying slowly at infinity and $α\in (0,1)$. Here, $h^L_α(s) = 1 - (1-s)^α L(\frac{1}{1-s}),$ as $s \to 1$. For a fixed $G$, the process $(h^L_α,G)$ explodes either for all $α\in (0,1)$ or for no $α\in (0,1)$, regardless of $L$. Next, we add contagious periods to all individuals and let their offspring survive only if their life-length is smaller than the contagious period of their mother: a forward process. An explosive process $(h^L_α,G)$, as above, stays explosive when adding a non-zero contagious period. We extend this setting to backward processes with contagious periods. Further, we consider processes with incubation periods during which an individual has already contracted the disease but is not able yet to infect her acquaintances. We let these incubation periods follow a distribution $I$. In the forward process $(h^L_α,G,I)_{f}$, every individual possesses an incubation period and only her offspring with life-time larger than this period survives. In the backward process $(h^L_α,G,I)_{b}$, individuals survive only if their life-time exceeds their own incubation period. These two processes are the content of the third main result that we establish: under a mild condition on $G$ and $I$, explosiveness of both $(h,G)$ and $(h,I)$ is necessary and sufficient for processes $(h^L_α,G,I)_{f}$ and $(h^L_α,G,I)_{b}$ to explode.

math.PR↗

arXiv2013

The effect of perception anisotropy on particle systems describing pedestrian flows in corridors

We consider a microscopic model (a system of self-propelled particles) to study the behaviour of a large group of pedestrians walking in a corridor. Our point of interest is the effect of anisotropic interactions on the global behaviour of the crowd. The anisotropy we have in mind reflects the fact that people do not perceive (i.e. see, hear, feel or smell) their environment equally well in all directions. The dynamics of the individuals in our model follow from a system of Newton-like equations in the overdamped limit. The instantaneous velocity is modelled in such a way that it accounts for the angle under which an individual perceives another individual. We investigate the effects of this perception anisotropy by means of simulations, very much in the spirit of molecular dynamics. We define a number of characteristic quantifiers (including the polarization index and Morisita index) that serve as measures for e.g. organization and clustering, and we use these indices to investigate the influence of anisotropy on the global behaviour of the crowd. The goal of the paper is to investigate the potentiality of this model; extensive statistical analysis of simulation data, or reproducing any specific real-life situation are beyond its scope.

physics.soc-ph↗