BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Eurandom - ECPv4.7.3//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:Eurandom
X-ORIGINAL-URL:https://www.eurandom.tue.nl
X-WR-CALDESC:Events for Eurandom
BEGIN:VEVENT
DTSTART;VALUE=DATE:20190325
DTEND;VALUE=DATE:20190330
DTSTAMP:20190119T040131
CREATED:20181107T130305Z
LAST-MODIFIED:20190111T090354Z
UID:2318-1553472000-1553903999@www.eurandom.tue.nl
SUMMARY:YEP XV "Information Diffusion on Random Networks"
DESCRIPTION:Summary\nThe "Information diffusion on random graphs" workshop is the 15th workshop in the ‘Young European Probabilists’ yearly workshops. \nDiffusion processes in networks manifests themselves in many real-life scenarios\, such as epidemic spreading\, viral marketing and power blackouts. This YEP workshop focuses information diffusion on networks. The phenomenon of information diffusion recently attracted vast attention across a wide range of research fields\, including mathematics\, physics\, computer science\, and social sciences. Therefore\, this YEP will focus not only on purely probabilistic aspects\, but also take an algorithmic and application perspective. The aim of the workshop is to bring together junior and senior researchers from probability and from other fields\, and to bridge the corresponding scientific communities. \nThe workshop will have three mini courses by internationally renowned researchers\, giving an opportunity to junior as well as senior attendants to learn about a new topic related to information diffusion. Other than that\, the workshop will consist of invited talks by junior and senior researchers. \n \nOrganizers\nRemco van der Hofstad (TU Eindhoven / Eurandom)\nNelly Litvak (University of Twente / TU Eindhoven)\nClara Stegehuis (TU Eindhoven) \n \nSpeakers\nTutorial speakers: \n\n\n\nFrank Ball\nUniversity of Nottingham\n\n\nMia Deijfen\nStockholm University\n\n\nRenaud Lambiotte\nUniversity of Oxford\n\n\n\n \n \nInvited speakers: \n\n\n\nClaudio Castellano\nSapienza\, Rome\n\n\nEric Cator\nRadboud University Nijmegen\n\n\nWei Chen\nMicrosoft Research Asia\n\n\nPetter Holme\nTokyo Institute of Technology\n\n\nLasse Leskelä\nAalto University\n\n\nMarton Karsai\nENS Lyon\n\n\nJuliá Komjathy\nTU Eindhoven\n\n\nNaoki Masuda\nUniversity of Bristol\n\n\nPeter Mörters\nCologna University\n\n\nDavid Sirl\nUniversity of Nottingham\n\n\nChi Tran\nUniversité des Sciences et Technologies de Lille\n\n\nDaniel Valesin\nUniversity of Groningen\n\n\nRose Yu\nNortheastern University\n\n\n\nContributed talks/posters\nDuring the conference we have a few slots available for contributed talks by participants. If you wish to apply for a contributed talk\, please send your title and abstract to koorn@eurandom.tue.nl.\nWe will also organize a poster session. If you would like to present a poster\, you can indicate this on the registration form. \n \nFunding possibilities\nFor this workshop\, we have reserved a limited budget that can be used to cover the local accommodation expenses of the speakers\, especially for young participants with limited resources. Speakers are encouraged to apply for the funding by sending an email to Patty Koorn at Koorn@eurandom.tue.nl. The deadline for the funding application is on\nJanuary 31. Funding applications will be treated on a first come first served basis. \n \nProgramme\nt.b.a. \n \nAbstracts\nFrank Ball (tutorial) \nEpidemics on networks\nThere has been considerable interest in the past two decades in models for the spread of epidemics on networks. The usual paradigm is that the population is described by an undirected random graph and disease can spread only along the edges of the graph. This mini-course gives an introduction to the analysis of SIR (susceptible-infective-recovered) epidemics on configuration model (and related) networks\, which is by far the most studied class of such epidemics. Topics covered include: \n\nbranching process approximation for the early stages of an epidemic\, which determines whether or not an epidemic with few initial infectives can become established and lead to a major outbreak;\nsusceptibility sets and the final outcome of a major outbreak;\neffective degree analysis of models\, which yields a functional central limit theorem (CLT) for the temporal behaviour and a CLT for the final outcome of a major epidemic;\nmodels with superimposed household structure\, a key component of human populations which can have a significant impact on disease dynamics;\nvaccination schemes\, including acquaintance vaccination which targets high-degree individuals.\n\n \nWei Chen (invited) \nInformation and Influence Propagation in Social Networks: Modeling and Influence Maximization\nInformation and influence propagation is a fundamental phenomenon in social networks that leads to many applications both for business and for public good\, such as viral marketing\, social recommendations\, rumor control\, epidemic prevention\, etc. In this talk\, I will survey the research area on information/influence diffusion dynamics and the influence maximization problem\, which is the problem of selecting a small number of seed nodes in a social network such that their influence spread is maximized. The talk will cover basic stochastic diffusion models\, algorithmic techniques for scalable influence maximization\, as well as some of my recent research work on influence-based centrality\, competitive and complementary influence diffusion\, etc. \n \nMia Deijfen (tutorial) \nCompeting growth on lattices and graphs\nCompeting first passage percolation describes the growth of two competing infections on an underlying graph structure. It was first studied on the Z^d-lattice. The main question is if the infection types can grow to occupy infinite parts of the lattice simultaneously\, the conjecture being that the answer is yes if and only if the infections grow with the same intensity. Recently\, the model has been analyzed on more heterogeneous graph structures\, where the degrees of the vertices can have an arbitrary distribution. In this case\, it turns out that also the degree distribution plays a role in determining the outcome of the competition. I will give a survey of existing results\, both on Z^d and on heterogeneous graphs\, and describe open problems. I will also describe related competition models such as the multitype contact process and models driven by moving particles. \n \nRenaud Lambiotte (tutorial) \nDiffusion and Communities in Networks\nThe presence of communities\, or clusters\, in networks is well known to affect diffusive processes. Conversely\, tracking the trajectories of random walkers on the graph can be used to uncover communities hidden in large graphs. During this tutorial\, I will review the relations between the two sides of the problem\, and present in detail community detection methods based on first-order and higher-order Markov models\, as well as methods allowing to uncover non-assortative communities in networks. \n \nPeter Gracar (contributed) \nSpread of infection by random walks - Multi-scale percolation along a Lipschitz surface\nA conductance graph on $\mathbb{Z}^d$ is a nearest-neighbor graph where all of the edges have positive weights assigned to them. We first consider a point process of particles on the nearest neighbour graph $(\mathbb{Z}^d\,E)$ and show some known results about the spread of infection between particles performing continuous time simple random walks. Next\, we extend consider the case of uniformly elliptic random graphs on $\mathbb{Z}^d$ and show that the infection spreads with positive speed also in this more general case. We show this by developing a general multi-scale percolation argument using a two-sided Lipschitz surface that can also be used to answer other questions of this nature. Joint work with Alexandre Stauffer. \n \nNaoki Masuda (invited) \nEpidemic processes on dynamically switching networks: Effects of commutator and concurrency\nEpidemic processes on temporally varying networks are complicated by complexity of both network structure and temporal dimensions. We analyse the susceptible-infected-susceptible (SIS) epidemic model on regularly switching networks\, where each contact network is used for a finite fixed duration before switching to another. First\, we analyse the epidemic threshold under a deterministic approximation called the individual-based approximation. We show that\, under this approximation\, temporality of networks lessens the epidemic threshold such that infections persist more easily in temporal networks than in their static counterparts. We further show that the commutator bracket of the adjacency matrices at different times is empirically a useful predictor of the impact of temporal networks on the epidemic threshold. The second topic is the effects of concurrency (i.e.\, the number of neighbours that a node has at a given time point) on the epidemic threshold in the stochastic SIS dynamics. For a particular switching network model\, we show that network dynamics can suppress epidemics (i.e.\, yield a higher epidemic threshold) when nodes' concurrency is low (where stochasticity effects are stronger) and can enhance epidemics when the concurrency is high. \n \nPeter Mörters (invited) \nMetastability of the contact process on evolving scale-free networks\nWe study the contact process in the regime of small infection rates on scale-free networks evolving by stationary dynamics. A parameter allows us to interpolate between slow (static) and fast (mean-field) network dynamics. For two paradigmatic classes of networks we investigate transitions between phases of fast and slow extinction and in the latter case we analyse the density of infected vertices in the metastable state. The talk is based on joint work with Emmanuel Jacob (ENS Lyon) and Amitai Linker (Universidad de Chile). \n \nGuilherme Reis (contributed) \nInteracting diffusions on random graphs\nWe consider systems of diffusion processes whose interactions are described by a graph. For example\, traditional mean-field interacting diffusions correspond to a complete interaction graph. In recent years some effort has been directed to understanding more general interactions. When the interaction graph is random\, in the particular case of the Erd\H{o}s-R\'{e}nyi random graph\, we show how the behavior of this particle system changes whether the mean degree of the Erd\"{o}s-R\'{e}nyi graph diverges to infinity or converges to a constant. When the mean degree converges to a constant we exploit a locality property of this system. Loosely speaking\, the locality property states that information does not propagate too fast over the graph for this kind of particle system. \n \nViktoria Vadon (contributed) \nPercolation on the Random Intersection Graph with Communities\nThe Random Intersection Graph with Communities (RIGC) models a network based on individuals and communities they are part of\, with two key features: each community has its arbitrary internal structure described by a small graph\, and communities are allowed to overlap. It generalizes the classical Random Intersection Graph (RIG) model\, and is constructed based on a Bipartite Configuration Model. We study percolation\, i.e.\, independent removal of edges\, as a simple model for a randomized information spread: we view the connected component of a vertex as the cluster this vertex is able to broadcast information to. We show that percolation on the RIGC\, in particular\, percolation on the classical RIG\, is (again) an RIGC with different parameters\, and prove that percolation on the RIGC exhibits a phase transition\, in terms of whether a linear-sized component persists. We may touch on robustness\, and why robustness of edge and vertex percolation behave differently. \n \nXiangying (Zoe) (contributed) \nThe Contact Process on Random Graphs and Galton-Watson Trees\nThe key to our investigation is an improved (and in a sense sharp) understanding of the survival time of the contact process on star graphs. Using these results\, we show that for the contact process on Galton-Watson trees\, when the offspring distribution (i) is subexponential the critical value for local survival $\lambda_2=0$ and (ii) when it is geometric($p$) we have $\lambda_2 \le C_p$\, where the $C_p$ are much smaller than previous estimates. We also study the critical value $\lambda_c(n)$ for ``prolonged persistence'' on graphs with $n$ vertices generated by the configuration model. In the case of power law and stretched exponential distributions where it is known $\lambda_c(n) \to 0$ we give estimates on the rate of convergence. Physicists tell us that $\lambda_c(n) \sim 1/\Lambda(n)$ where $\Lambda(n)$ is the maximum eigenvalue of the adjacency matrix. Our results show that this is not correct. \n \nDong Yao (contributed)\nThe symbiotic contact process\nWe consider a contact process on $\ZZ^d$ with two species that interact in a symbiotic manner. Each site can either be vacant or host individuals of species A and/or B. Multiple occupancy by the same species at a single site is prohibited. Symbiosis is represented by a reduced death rate $\mu \in [0\,1) $. If only one specie is present at a site then that particle dies with rate 1 but if both species are present then the death rate is reduced to $\mu$ for the two particles at that site. We prove that the critical infection rate $\lambda_c(\mu)$ for weak survival is of order $\sqrt{\mu}$\, which coincides with the mean field calculation. We also investigate the nature of the phase transition. We show that in dimension $d=1$ the survival of the system is through oriented percolation. We also show that\, for all dimensions\, the phase transition is continuous and $\lambda_c(\mu)$ is 1 (regardless of the value of $\mu$)\, if we let particles move around with a rate going to infinity. The talk is based on ongoing work with Rick Durrett. \n \nXiu-Xiu Zhan (contributed) \nInformation Diffusion Backbones in Temporal Networks\nInformation diffusion on a temporal network can be modeled by viral spreading processes such as the Susceptible-Infected (SI) spreading process. An infected node meaning that the node possesses the information could spread the information to a Susceptible node with a given spreading probability β whenever a contact happens between the two nodes. Progress has been made in the understanding of how temporal network features and the choice of the source node affect the prevalence\, i.e. the percentage of nodes reached by the information. In this work\, we explore further: which node pairs are likely to contribute to the actual diffusion of information\, i.e. appear in a diffusion trajectory? How is this related to the local temporal connection features of the node pair? Such deep understanding of the role of node pairs is crucial to explain and control the prevalence of information spread. First\, we propose the construction of an information diffusion backbone G_B (β) for an SI spreading process with an infection probability β on a temporal network. The backbone is a weighted network where the weight of each node pair indicates how likely the node pair contributes to a diffusion process starting from an arbitrary node. Second\, we investigate the relation between the backbones with different infection probabilities on a temporal network. We find that the backbone topologies obtained for low and high infection probabilities approach the backbone G_B (β→0) and G_B (β=1)\, respectively. The backbone G_B (β→0) equals the integrated weighted network\, where the weight of a node pair counts the total number of contacts in between\, a local temporal connection feature. Finally\, we discover a local connection feature among many other features that could well predict which node pairs are likely to appear in G_B (β=1)\, whose computation complexity is high. This local feature encodes the time that each contact occurs\, pointing out the importance of temporal features in determining the role of node pairs in a dynamic process beyond the features of the integrated network. \n \nRegistration\nLink to the online registration form: REGISTRATION \nPractical Information\nInformation on travel\, location etc. : INFORMATION \n
URL:https://www.eurandom.tue.nl/event/yep-xv-information-diffusion-on-random-networks/
LOCATION:Eurandom\, Metaforum\, Eindhoven\, Netherlands
CATEGORIES:YEP
END:VEVENT
END:VCALENDAR