2017 spring

23 Marc
István Miklós (Rényi Institute): Exact sampling of graphs with prescribed degree correlations
Abstract: Many realworld networks exhibit correlations between the node degrees. For instance, in social networks nodes tend to connect to nodes of similar degree. Conversely, in biological and technological networks, highdegree nodes tend to be linked with lowdegree nodes. Degree correlations also affect the dynamics of processes supported by a network structure, such as the spread of opinions or epidemics. The proper modelling of these systems, i.e., without uncontrolled biases, requires the sampling of networks with a specified set of constraints. We present a solution to the sampling problem when the constraints imposed are the degree correlations. Furthermore we would like to give a brief survey of our current knowledge of generating such random networks.

9 Marc
Federico Musciotto(University of Palermo and CEU): Timely evolution and core of communities of statistically validated projections of bipartite networks
Abstract: Many complex systems are naturally organized in bipartite networks, i.e. networks with two disjoint sets of node, usually characterized by high heterogeneity. Although useful information can be extracted from these networks, their projections are not always stable and robust against errors in the data or other sources of noise. A solution to this issue is based on statistically validated networks, which reduce the original system by considering only the links which are statistically significant. In this way, cores of the real communities of the system are obtained with a high level of precision. Moreover, the methodology of SVN can be extended in order to track the evolution in time of the communities of a system. In this talk I will show the results of these methods on a dataset of stock market Finnish investors, characterized by high heterogeneity with respect to trading activity. This work is the result of a collaboration with Luca Marotta, Jyrki Piilo and Rosario N. Mantegna.
2016 fall

20 Dec
András Bóta (The University of New South Wales, Australia): Learning infection processes
Abstract: Infection models can be used to model the spread of disease, information, behaviour and many other things through a network composed of connected nodes. One of the common challenges arising in the application of infection models is the lack of available transmission probabilities. The task of inverse infection is the systematic estimation of these values. Several methods have been proposed recently for solving this task. A common property of these approaches is that they make many specialized assumptions about the problem they are trying to solve. In contrast the method we discuss in this talk gives a general framework for inverse infection tasks and offers much needed flexibility allowing the method to be used with a variety of reallife applications. In the second part of the talk we are going to discuss a specific application of the generalized inverse infection model. Modern transportation infrastructure allows quick and efficient travel between distant parts of the world, but unfortunately also offers many ways to transfer diseases between regions. Recent examples of global outbreaks include the MERS, SARS and swine flu epidemics and most recently the 2015 Zika fever outbreak of Brazil. A critical component of outbreak control is the identification of disease spreading from region to region. We will show how the proposed inverse infection model can be used to 1. accurately model the Zika virus outbreak on the country level, 2. estimate travel risk between regions.

17 Nov
Brys Zoltán (LAM, BMETMIT): Hálózatkutatás a népegészségügy és a gyógyszeripari marketing területén. Esetismertetés és áttekintés. A gyógyszerfelírások, a pattanások és fejfájás valóban "szociálisan fertőznek"?
Kivonat: A gráfelmélet a fertőző betegségek terjedésének modellezésében régóta használt módszer és fontos eredményeket tárt fel (Epstein, Balcan, Lilrejos stb.). A fehérjefehérje interakciós hálózat elemzése pedig a bioscience szerves részévé vált (genetika, epigenetika, gyógyszerfejlesztés stb.) A képalkotó eljárások és az optogenetika terjedésével a neurológiában is egyre inkább használt a gráfelmélet. A népegészségügy területén nagy vitát váltott ki Christakis és Fowler 2007es felvetése, amely szerint a noncommunicable diseases egyik jelentős okaként számon tartott elhízás és dohányzás "szociálisan fertőz". A felvetés jelentős kritikát kapott, főként statisztikai (Lyons) és szociálpszichológiai/szociológiai (Buda) területen. CohenCole és mtsai (Yale) szellemes közleményében hasonló módszertannal bizonyította, hogy Christakisék logikájával hálózatos hatások érvényesülnek a pattanások és fejfájások terjedésében is. A hálózatelemzés ígérete, hogy egyfajta "theory of everything", mindent magyarázó "csodamódszer" lesz nem látszik beigazolódni. Sajnos az egyes hálózatelemzéssel dolgozó sztárszerzők a "csodamódszer" birtokában könnyen és észrevétlenül (és biztosan nem tudatosan) negligálták a társadalomtudományokban és a társtudományokban felhalmozott tudásvagyont (pl. szociológia, addiktológia stb.). Az előadó egy valós gyógyszeripari Clinical Opion Leader detektálási eset ismertetése kapcsán próbálja bemutatni, hogy a népegészségügyhöz hasonló folyamat látszik a gyakorlati tudományok területén, különösen a marketing az adatelemzés területére. Az előadó személyes meglátása szerint a hálózatelemzés hypefázisa lassan lejár és a módszer valódi értéke lassan kibontakozni, ami jelentősen kisebb, mint az ígéret volt, de így is rendkívül jelentős.

3 Nov
Ivan Fekete(Semmelweis University and Linkgroup): Computational prediction personalized of drug combinations
Abstract: Numerous methods and biological networks have been used to model intracellular signal transduction, but so far the clinical applicability of these approaches remained rather limited. Here we describe a novel system (Turbine, http://turbine.hu) for the reliable simulation of intracellular signaling which required, first, to utilize a large, fully dynamically reviewed, manually curated network of major human signaling pathways and their transcriptional regulatory mechanisms, and second, to run ensembles of simulations and extract the resulting attractors  steady states  of the system. The software finds the attractors of the signaling network, and correlates the attractors’ activity patterns with the activity of biological processes, like apoptosis or proliferation. By combining a large set of steady states, we were able to map the cellular attractor landscapes, which varied depending on the presence of different physiological ligands, available membrane receptors, or  most importantly regarding the clinical applicability  mutated proteins. This approach, combined with additional omics data layers (such as cancer genomic and transcriptomic profiles) and artificial intelligence systems made Turbine able to predict potential mono or combination drug therapies on a personalized basis.

20 Oct
Laszlo Kovacs (University of West Hungary, Department of Applied Lingusitics): Linguistic networks
Abstract: Network structures are not new in linguistics: since the 1960’s networks are used to explain linguistic phenomenon. In the first part of the lecture a broad picture of linguistic network research will be given, showing where networked structures in language exist. The main focus of the lecture is on the networked structures of the mind: it will be shown on the example of the Hungarian database "ConnectYourMind" which structures in the mental lexicon ("dictionary of mind") exist. The presented results are joint works done with Andras Bota, Laszlo Hajdu and Miklos Kresz (University of Szeged) and with Peter Pollner and Katalin Orosz (Eotvos University).

6 Oct
Gyorgy Turan (University of Szeged and University of Illinois at Chicago): Betweenness centrality
Abstract: The betweenness centrality of a vertex in a network measures the number of shortest paths containing the vertex. The local version considers only shortest paths of bounded length. We review related results, and then discuss the behavior of the local version for trees, including worstcase and scalefree random trees. Joint work with Ben Fish and Rahul Kushwaha.
2016 spring

12 May
Nandor Poka, (research assistant (Department of Applied Informatics) and doctoral candidate (Doctoral School of Multidisciplinary Medical Science) University of Szeged): Combinatorial Scientific Computing and (Computational) Biology: When the demand meets the offer
Abstract: With the rise of the BigData and HighPerformance Computing era, came the flourishing of many fields of science, including Combinatorial Scientific Computing. Both the business and scientific community produces such vast amounts of data, that new technologies and algorithms are required to analyze them. To efficiently utilize highly parallel computers or clusters, tasks must be decomposed and the data must be partitioned, and these involve graph algorithms themselves. Combinatorial Scientific Computing methods have been used for eg. in the aforementioned load distribution, automatic differentiation, statistical physics, and other enigmatic areas. However more “lifelike” applications of these methods were and are put to use in other fields of science like modern biology. In our talk we will present various biological problems (both “old” and more recent) that can be fairly easily represented with mathematical objects such as graphs and matrices. Our main focus of these problems will be the present and future challenges of NextGeneration Sequencing, and how to tackle them using Data Science and Combinatorial Scientific Computing.

28 April
Miklos Kresz (Department of Applied Informatics, University of Szeged; joint work with Andras Bota and Andras Pluhar): Dynamic network mining and business intelligence
Abstract: During the last decade social network analysis and mining became a key research area. Apart from the obvious applications in online social network services, the field plays central role in many classical business intelligence tasks such as customer attrition, risk analysis and campaign management. In order to capture the characteristics of the above problems, the dynamics of the corresponding network processes and that of the changes in the network structure needs to be studied. In this talk we will consider two relevant problems. Dynamic community detection is an algorithmic tool for the analysis of the lifetime of communities in real graphs. The study of infection processes in networks pose several algorithmic and modelling questions such as maximizing the spread of influence or approximating the real infection values. In addition to review applied models and methods, in the talk reallife applications will be also presented.

7 April
Bogdan Zavalnij(University of Pecs): Mathematical modeling of various problems by graphs and solving them with clique search
Abstract: In our talk we would like to demonstrate the modeling powers of graphs. We will demonstrate problems from different fields and show that they can be reformulated as graphs. The solution of these problems will be transformed to usual graph problems as different coloring and clique problems. The problems included in the talk will be from simple games, scheduling, stock exchange, drug design and even graph problems themselves  including hypergraph coloring problems. We would like to show some useful techniques and also some common methods for such reformulations that could prove useful.

24 March
Christian Bongirono (University of Palermo): The dualprojection approach for bipartite networks using statistical link validation
Abstract: Bipartite complex networks are usually analyzed by projecting the two disjoint set of nodes into two networks and then using the standard techniques for them separately. Since complex systems are often very heterogeneous that makes very difficult to distinguish links of the projected network that are just reflecting system’s heterogeneity from links relevant to unveil the properties of the system. To avoid this problem it has been developed a methodology for onemode projections of bipartite networks using an unsupervised statistical link validation. In order to study the efficiency of the method we investigate the community structure of the projected network using both a simple projection and the statistically validated projection on various synthetic benchmarks and real networks. In all these cases the link validating filtering procedure necessarily increases the precision and suggested to use, even if considering the drawback that it decreases the level of accuracy in certain situations.

10 March
Christian Bongirono (University of Palermo): Statistical Regularity in the Air Traffic flow
Abstract: The aircraft trajectories are compounded by a sequence of spatial fixed points (NVP) that typically diverge from the best path route. This infrastructure allows the air traffic controllers to direct the air traffic flow on standard air ways, and focuses their attention to a few numbers of special NVPs where the routes converge. As a drawback the not optimal routes force the air traffic controller to modify, where is possible, the routes to enhance the air traffic flow. The aim of the first part of the talk is to highlight the behaviour of the air traffic controller respect to a network optimization operation named direct by the observation of stylized facts both at the global trajectory level and at the local navigation point level. In the second part of the talk will be discussed how an Agent Based Model could help us in understanding the transition from the current NVP based network to the future new SESAR scenario, where the aircraft will be allowed to follow a freeroute path.
2015 fall

26 November
Gabor Berend (Institute of Informatics): Learning the transition matrix for weighted PageRank
Abstract: The PageRank algorithm is a widely known and well understood approach to assign importance scores to nodes in networks. This highly applicable approach relies on the assumption that all the connections of a node has equal importance (i.e. inversely proportional to their outdegree). The talk will introduce approaches that aim at determining the strength of the connections between pairs of nodes given that the PageRank scores (or some relative importance scores) of the nodes within the graph is assumed to be known in advance. After presenting these models, both quantitative and qualitative results will be presented on synthetic and real world networks (e.g. citation and collaboration networks, the English and Hungarian Wikipedia and networks generated from language usage).

19 November
Gabor London (MTASZTE Stereochemistry Research Group): Accomplishments and challenges of networks science in chemistry
Abstract: Approaching chemical problems with the tools of network science has proved very fruitful in recent years. The network approach has provided us with new ways of designing drugs, optimizing multistep chemical syntheses or fighting terrorism. Most of these successes, however, are based on the analyses of available data. One of the big challenges chemists are facing whether it is possible to implement this databased knowledge for the creation of instructable molecular networks that can be models for early (molecular) evolution or able to perform complex (synthetic) tasks. The talk will discuss both the databased achievements and the recent attempts towards creating “molecular ecosystems”.

5 November
Istvan Kiss (SZTE Knowledge Management Research Center): Communities and central nodes in the network of the mobile inventors in the United States
Abstract: IPintensive industries accounted for about 3340% of the gross domestic product in western economies. These intellectual properties are mostly embodied in patents and trademarks. Highest stake of the expenditures in the mentioned sector goes for the wages of whitecollar workers what underlines the importance of the intellectual capital and knowledge in the creation and development of IP portfolio. We investigated the flow of knowledge as a crucial resource among organizations by analysis of patent documents from the United States. In our network organizations are the nodes and the mobility of researchers among them are the edges. This graph can be considered as the informal innovation network of US organizations, where firms, universities or governmental institutions competing for knowledge and recombine their innovation capacity through the mobility of inventors.

22 October
Tamas Vinko (Institute of Informatics): Network models for BitTorrent communities
Abstract: BItTorrent communities are content sharing systems using peertopeer technology. From mathematical point of view these communities can be modeled with graphs of particular structure. For example, there is a straightforward bipartite representation, where we take the users and the files as vertices and the edges represent supply and demand. Using this simple representation one can already consider interesting optimization problems. Moreover, there is a richer graph representation which leads to a flow network. Hence, one can start thinking about applying traditional flow algorithms and their meaning in this particular context.

8 October
Andras London (Institute of Informatics): Statistically validated projections of bipartite networks
Abstract: Bipartite networks naturally appear in from social to biological systems, examples include, among many others, the actorsmovies network, artistsmusic network, scientistsresearch papers cooperation network, network of sexual contacts, diseasesgenes network, plantspollinators mutualistic networks, banksfirms money transfer networks and words cooccurrence networks. Many properties of these networks typically investigated by constructing and analyzing a projected network on one of the two sets of the original network. When one constructs a projected network of nodes only from one set, the original network's heterogeneity (e.g. the heterogeneous degree distribution) makes difficult to determine those links that presence in the projected network cannot be explained by random cooccurrence of their neighbors in the original network. In this talk we present a method based on statistical validation to overcome this problem and show some possible applications on real bipartite systems.