Network Science for Multidimensional Data Analysis
The coexistence of multiple types of interactions within social, technological and biological networks has moved the focus of the statistical physics of complex systems towards their description as a set of subsystems organized as layers of connectivity. This approach has unveiled that the multilayer nature of complex systems has strong influence in the emergence of collective states and their critical properties, setting a novel paradigm in the past decade.
However, real-world systems are not only multilayer in their nature: they exhibit, simultaneously, a complex organization across multiple scales in their topology, dynamics and function. Recently, many approaches have been proposed to model higher-order interactions different from multiplexity: from simplicial complexes to memory in network flows, from latent topological geometry to multiresolution methods to unravel the hidden geometry of network-driven processes.
All those methods provide powerful tools to analyze complex systems and to unravel the effects of hierarchies from different points of view. However, empirical networks often exhibit multiscale spatio-temporal organization, multilayer relationships and non-trivial geometry.
Aim of MIX-NEXT. The aim of this Satellite meeting is to balance the contribution of well established leading experts and rising young researchers to review the recent advances in those research fields, with the aim of triggering and igniting new discussions on theoretical and computational solutions required to build a more comprehensive set of tools integrating different perspectives into one, coherent and self-consistent, framework for modeling and analysis of complex networks.
The list of topics that we aim to cover at the conference is the following:
- Mathematical properties of multiscale, multilayer and (hidden) geometric structures
- Empirical measurements for multiscale, multilayer and higher-order networks
- Applications of such models to biological, social, technological and urban systems
University College London
Ben-Gurion University of the Nagev
University of Pennsylvania
Central European University
|Rita Maria del Rio-Chanona
University of Oxford
|Guilherme Ferraz de Arruda
Universidad de Zaragoza
Universitat Rovira i Virgili
University of Washington
|16 November 2020||Early Registration Deadline|
|4th December 2020||yrCSS Warm-up|
|7-11 December 2020||Conference|
|9 December 2020||Satellite event|
|9 Dec: 13:00||
Manlio De Domenico
Opening and Chair
Jesus Gomez-Gardenes, Universidad de Zaragoza
Modeling and analysis of COVID-19 diffusion by integrating multiple interactions and mobility patterns
The spread of COVID-19 is posing an unprecedented threat to health systems worldwide. The fast propagation of the disease combined with the existence of covert contagions by asymptomatic individuals make the controlling of this disease particularly challenging. Here, we propose a metapopulation model that integrates into a single framework the different mobility and interaction patterns that coexist in our societies. These mixing and mobility patterns are the main drivers behind SARS-CoV-2 diffusion and the impact of COVID-19 impact on our health systems. This tailored epidemic model allows us to monitor and explain the advance of the disease and to find an analytical expression for the effective reproduction number, R, as a function of mobility restrictions and confinement measures. The expression for R is an extremely useful tool to design containment policies that are able to suppress the epidemics allowing us to determine the precise reduction of mobility needed to bend the curve of epidemic incidence.
Elsa Arcaute, University College London
Hierarchies and spillovers in systems of systems
In this talk we will explore how to analyse urban systems integrating the different scales, and the different systems. Dynamical processes take place at a different pace in each of the different systems, and these need to be coupled. The hierarchical organisation emerging from these systems can help in coupling the systems at the "right" level and observe the effects at different scales.
Federico Battiston, Central European University
Beyond pairwise interactions: structure and dynamics of higher-order systems
The complexity of many biological, social and technological systems stems from the richness of the interactions among their units. Over the past decades, a variety of complex systems has been successfully described as networks whose interacting pairs of nodes are connected by links. Yet, from human communications to chemical reactions and ecological systems, interactions can often occur in groups of three or more nodes and cannot be described simply in terms of dyads. Until recently little attention has been devoted to the higher-order architecture of real complex systems. However, a mounting body of evidence is showing that taking the higher-order structure of these systems into account can enhance our modeling capacities and help us understand and predict their dynamical behavior. In this talk I will introduce the different frameworks to represent higher-order systems, such as simplicial complexes and hypergraphs, and discuss new emergent phenomena and collective behavior characterizing dynamical processes when extended beyond pairwise interactions.
Rita Maria del Rio-Chanona, University of Oxford
he multilayer nature of the economy and global financial contagions
As illustrated by the 2008 global financial crisis, the financial distress of one country can trigger financial distress in other countries. We examine the problem of identifying such “systemically important” countries (i.e., countries whose financial distress can trigger further distress), which is important for assessing global financial stability. Using data on bilateral financial positions that are split by asset type, we build a multiplex global financial network in which nodes represent countries, edges encode cross-country financial assets of various types, and layers represent asset types. We develop a multiplex threshold model of financial contagions in which a shock can propagate either within a layer or between layers. We find that the number of systemically important countries can be twice as large when we take into account the heterogeneity of financial exposures (i.e., when using a multiplex network) than in a contagion on an associated aggregate global financial network (i.e., on a monolayer network), as is often examined in other studies. Our analysis suggests that accounting for both intralayer and interlayer propagation of contagions in a multiplex structure of financial assets is important for understanding interconnected financial systems of countries. At the end of the talk, I'll discuss the importance of considering the multilayer nature of the economy for understanding the economic impact of the COVID-19 pandemic.
Yosef Ashkenazy, Ben-Gurion University of the Negev
Climate change through climate network
The climate system plays a central part in human life and as such was and is investigated intensively by many research groups. Yet, different climate phenomena are still not fully understood. In the talk we present three examples demonstrating how climate network analysis uncover climate phenomena and improve our understating regarding the climate system. The examples are: (1) Dominant imprint of Rossby waves in the climate network, (2) Significant impact of Rossby waves on air pollution detected by network analysis, and (3) Climate network percolation reveals the expansion and weakening of the tropical component under global warming.
Danielle Bassett, University of Pennsylvania
Deep phenotyping: An opportunity to understand multilayer networks in the human brain
Empirical studies of human anatomy, physiology, and behavior typically fall along a spectrum from broad studies of many people with sparse data per person to deep studies of a few people with dense data per person. The latter studies are typically referred to as deep phenotyping studies, where each person is studied along multiple dimensions, with several data types collected simultaneously. Deep phenotyping offers a notable opportunity for the study of phenomena that arise from multilayer network systems in the biology and psychology of the human. While building networks from multiple data types can be easy, knowing precisely what questions to ask of them can be more difficult. Here I will posit two particularly useful goals of building such networks: (i) to explain observations in a layer late in the causal chain from observations in a layer early in the causal chain, and (ii) to build generative models that can bolster inferences from sparsely sampled patient data. I will work through a case study from human neuroimaging to illustrate these goals. Collectively the work underscores the importance of combining multiple data modalities in deep phenotyping studies, and highlights important challenges in that process.
Guilherme Ferraz de Arruda, ISI Foundation
Social contagion models on hypergraphs
Our understanding of the dynamics of complex networked systems has increased significantly in the last two decades. However, most of our knowledge is built upon assuming pairwise relations among the system's components. This is often an oversimplification, for instance, in social interactions that frequently occur within groups. To overcome this limitation, here we study the dynamics of social contagion on hypergraphs. We develop an analytical framework and provide numerical results for arbitrary hypergraphs, which we also support with Monte Carlo simulations. Our analyses show that the model has a vast parameter space, with first- and second-order transitions, bistability, and hysteresis. Phenomenologically, we also extend the concept of latent heat to social contexts, which might help understand oscillatory social behaviors. Our work unfolds the research line of higher-order models and the analytical treatment of hypergraphs, posing new questions, and paving the way for modeling dynamical processes on higher-order structures.
Marta Sales-Pardo, Universitat Rovira i Virgili
From recommendation to obtaining models from data using inference: Is the data always enough?
I will talk about two different problems that share a common behavior: a transition between a desired outcome inference and a less desirable one. My first example example will be that of recommendation systems (or bipartite graphs with multi-valued edges) and the use of attributes (such as gender of a user and the genre of a movie) to increase prediction accuracy, My second example will be that of obtaining models from data using a Bayesian inference framework in the presence of noise. I will discuss how adata attributes are not always useful to make recommendations in the same way that increasing the noise in the data will prevent us from finding the model that generated the data. I will also discuss that if we increase the importnce of the attributes or the noise we observe a transition between a regime in which we only see the data and a regime in which we only see the attributes or models that are comaptibles with noise.
Alice Schwarze, University of Washington
Motifs for processes on networks
The study of motifs in networks can help researchers uncover links between structure and function of networks in biology, the sociology, economics, and many other areas. Empirical studies of networks have identified feedback loops, feedforward loops, and several other small structures as "motifs" that occur frequently in real-world networks and may contribute by various mechanisms to important functions these systems. However, the mechanisms are unknown for many of these motifs. We propose to distinguish between "structure motifs" (i.e., graphlets) in networks and "process motifs" (which we define as structured sets of walks) on networks and consider process motifs as building blocks of processes on networks. Using the covariances and correlations in a multivariate Ornstein--Uhlenbeck process on a network as examples, we demonstrate that the distinction between structure motifs and process motifs makes it possible to gain quantitative insights into mechanisms that contribute to important functions of dynamical systems on networks.
Filippo Radicchi, Indiana University
Decoding communities in networks
In this talk, I will present an interpretation of the problem of defining and identifying communities in networks as a classical communication task over a noisy channel: memberships of nodes are information bits erased by the channel, edges and non-edges in the network are parity bits introduced by the encoder but degraded through the channel, and a community identification algorithm is a decoder. The interpretation is perfectly equivalent to the one at the basis of well-known statistical inference algorithms for community detection. The only difference in the interpretation is that a noisy channel replaces a stochastic network model. However, the different perspective gives the opportunity to take advantage of the rich set of tools of coding theory to generate novel insights on the problem of community detection. I will illustrate two main applications of standard coding-theoretical methods to community detection. First, I will take advantage of a state-of-the-art decoding technique to generate a family of quasi-optimal community detection algorithms. Second and more important, I will show that the Shannon's noisy-channel coding theorem can be invoked to establish a lower bound, here named as decodability bound, for the maximum amount of noise tolerable by an ideal decoder to achieve perfect detection of communities. When computed for well-established synthetic benchmarks, the decodability bound explains accurately the performance achieved by the best community detection algorithms existing on the market, telling us that only little room for their improvement is still potentially left.