arXiv subjects

Vito Latora

Publications and source records attributed to Vito Latora.

At least 37 records · Page 2Linked to original sources

arXiv2023

Hyperedge overlap drives explosive collective behaviors in systems with higher-order interactions

Recent studies have shown that novel collective behaviors emerge in complex systems due to the presence of higher-order interactions. However, how the collective behavior of a system is influenced by the microscopic organization of its higher-order interactions remains still unexplored. In this Letter, we introduce a way to quantify the overlap among the hyperedges of a higher-order network, and we show that real-world systems exhibit different levels of hyperedge overlap. We then study models of complex contagion and synchronization of phase oscillators, finding that hyperedge overlap plays a universal role in determining the collective dynamics of very different systems. Our results demostrate that the presence of higher-order interactions alone does not guarantee abrupt transitions. Rather, explosivity and bistability require a microscopic organization of the structure with a low value of hyperedge overlap.

physics.soc-ph↗

arXiv2023

Collapse transition in epidemic spreading subject to detection with limited resources

Compartmental models are the most widely used framework for modeling infectious diseases. These models have been continuously refined to incorporate all the realistic mechanisms that can shape the course of an epidemic outbreak. Building on a compartmental model that accounts for early detection and isolation of infectious individuals through testing, in this article we focus on the viability of detection processes under limited availability of testing resources, and we study how the latter impacts on the detection rate. Our results show that, in addition to the well-known epidemic transition at ${\mathcal{R}}_0=1$, a second transition occurs at ${\mathcal{R}}^*_0>1$ pinpointing the collapse of the detection system and, as a consequence, the switch from a regime of mitigation to a regime in which the pathogen spreads freely. We characterize the epidemic phase diagram of the model as a function of the relevant control parameters: the basic reproduction number, the maximum detection capacity of the system, and the fraction of individuals in shelter. Our analysis thus provides a valuable tool for estimating the detection resources and the level of confinement needed to face epidemic outbreaks.

physics.soc-ph↗

arXiv2023

Reconstructing higher-order interactions in coupled dynamical systems

Higher-order interactions play a key role for the stability and function of a complex system. However, how to identify them is still an open problem. Here, we propose a method to fully reconstruct the structural connectivity of a system of coupled dynamical units, identifying both pairwise and higher-order interactions from the system time evolution. Our method works for any dynamics, and allows the reconstruction of both hypergraphs and simplicial complexes, either undirected or directed, unweighted or weighted. With two concrete applications, we show how the method can help understanding the ecosystemic complexity of bacterial systems, or the microscopic mechanisms of interaction underlying coupled chaotic oscillators.

physics.soc-ph↗

arXiv2023

Complex contagion in social systems with distrust

Social systems are characterized by the presence of group interactions and by the existence of both trust and distrust relations. Although there is a wide literature on signed social networks, where positive signs associated to the links indicate trust, friendship, agreement, while negative signs represent distrust, antagonism, and disagreement, very little is known about the effect that signed interactions can have on the spreading of social behaviors when higher-order interactions are taken into account. In this paper we focus on processes of complex contagion, such as the adoption of social norms, where exposure to multiple sources is needed for the contagion to occur. Complex contagion has been recently modeled by higher-order networks, such as simplicial complexes, which allow transmission to happen not only through the links connecting pair of nodes, but also in group interactions, namely over simplices of dimension larger or equal than two. Here, we introduce a model of complex contagion on signed simplicial complexes, and we investigate the role played by trust and distrust on the dynamics of a social contagion process. The presence of higher-order signed structures in our model naturally induces new infection and recovery mechanisms. Through numerical simulations and analytical results in the mean-field approximation, we show how distrust determines the way the system moves from a state where no individuals adopt the social behavior, to a state where a finite fraction of the population actively spreads it. Interestingly, the fraction of spreading individuals displays a non-monotonic dependence on the average number of connections between individuals. We then investigate how social balance affects social contagion, finding that balanced triads either promote or impede contagion based on the relative abundance of fully trusted relations.

physics.soc-ph↗

arXiv2023

Explosive cooperation in social dilemmas on higher-order networks

Understanding how cooperative behaviours can emerge from competitive interactions is an open problem in biology and social sciences. While interactions are usually modelled as pairwise networks, the units of many real-world systems can also interact in groups of three or more. Here, we introduce a general framework to extend pairwise games to higher-order networks. By studying social dilemmas on hypergraphs with a tunable structure, we find an explosive transition to cooperation triggered by a critical number of higher-order games. The associated bistable regime implies that an initial critical mass of cooperators is also required for the emergence of prosocial behavior. Our results show that higher-order interactions provide a novel explanation for the survival of cooperation.

physics.soc-ph↗

arXiv2023

Higher-order correlations reveal complex memory in temporal hypergraphs

Many real-world complex systems are characterized by interactions in groups that change in time. Current temporal network approaches, however, are unable to describe group dynamics, as they are based on pairwise interactions only. Here, we use time-varying hypergraphs to describe such systems, and we introduce a framework based on higher-order correlations to characterize their temporal organization. We analyze various social systems, finding that groups of different sizes have typical patterns of long-range temporal correlations. Moreover, our method reveals the presence of non-trivial temporal interdependencies between different group sizes. We introduce a model of temporal hypergraphs with non-Markovian group interactions, which reveals complex memory as a fundamental mechanism underlying the pattern in the data.

physics.soc-ph↗

arXiv2023

Maximal dispersion of adaptive random walks

Maximum entropy random walks (MERWs) are maximally dispersing and play a key role in optimizing information spreading in various contexts. However, building MERWs comes at the cost of knowing beforehand the global structure of the network, a requirement that makes them totally inadequate in real case scenarios. Here, we propose an adaptive random walk (ARW), which instead maximizes dispersion by updating its transition rule on the local information collected while exploring the network. We show how to derive ARW via a large-deviation representation of MERW and study its dynamics on synthetic and real world networks.

cond-mat.stat-mech↗

arXiv2022

Evolutionary game model of group choice dilemmas on hypergraphs

We introduce an evolutionary game on hypergraphs in which decisions between a risky alternative and a safe one are taken in social groups of different sizes. The model naturally reproduces choice shifts, namely the differences between the preference of individual decision makers and the consensual choice of a group, that have been empirically observed in choice dilemmas. In particular, a deviation from the Nash equilibrium towards the risky strategy occurs when the dynamics takes place on heterogeneous hypergraphs. These results can explain the emergence of irrational herding and radical behaviours in social groups.

physics.soc-ph↗

arXiv2022

Model of spatial competition on discrete markets

We propose a dynamical model of price formation on a spatial market where sellers and buyers are placed on the nodes of a graph, and the distribution of the buyers depends on the positions and prices of the sellers. We find that, depending on the positions of the sellers and on the level of information available, the price dynamics of our model can either converge to fixed prices, or produce cycles of different amplitudes and periods. We show how to measure the strength of competition in a spatial network by extracting the exponent of the scaling of the prices with the size of the system. As an application, we characterize the different level of competition in street networks of real cities across the globe. Finally, using the model dynamics we can define a novel measure of node centrality, which quantifies the relevance of a node in a competitive market.

physics.soc-ph↗

arXiv2022

Dual communities in spatial networks

Both human-made and natural supply systems, such as power grids and leaf venation networks, are built to operate reliably under changing external conditions. Many of these spatial networks exhibit community structures. Here, we show that a relatively strong connectivity between the parts of a network can be used to define a different class of communities: dual communities. We demonstrate that traditional and dual communities emerge naturally as two different phases of optimized network structures that are shaped by fluctuations and that they suppress failure spreading, which underlines their importance in understanding the shape of real-world supply networks.

physics.soc-ph↗

arXiv2022

Turing patterns in systems with high-order interactions

Turing theory of pattern formation is among the most popular theoretical means to account for the variety of spatio-temporal structures observed in Nature and, for this reason, finds applications in many different fields. While Turing patterns have been thoroughly investigated on continuous support and on networks, only a few attempts have been made towards their characterization in systems with higher-order interactions. In this paper, we propose a way to include group interactions in reaction-diffusion systems, and we study their effects on the formation of Turing patterns. To achieve this goal, we rewrite the problem originally studied by Turing in a general form that accounts for a microscropic description of interactions of any order in the form of a hypergraph, and we prove that the interplay between the different orders of interaction may either enhance or repress the emergence of Turing patterns. Our results shed light on the mechanisms of pattern-formation in systems with many-body interactions and pave the way for further extensions of Turing original framework.

nlin.PS↗

arXiv2022

Deep Neural Networks as Complex Networks

Deep Neural Networks are, from a physical perspective, graphs whose `links` and `vertices` iteratively process data and solve tasks sub-optimally. We use Complex Network Theory (CNT) to represents Deep Neural Networks (DNNs) as directed weighted graphs: within this framework, we introduce metrics to study DNNs as dynamical systems, with a granularity that spans from weights to layers, including neurons. CNT discriminates networks that differ in the number of parameters and neurons, the type of hidden layers and activations, and the objective task. We further show that our metrics discriminate low vs. high performing networks. CNT is a comprehensive method to reason about DNNs and a complementary approach to explain a model's behavior that is physically grounded to networks theory and goes beyond the well-studied input-output relation.

cs.LG↗

arXiv2022

Hyper-diffusion on multiplex networks

Multiplex networks describe systems whose interactions can be of different nature, and are fundamental to understand complexity of networks beyond the framework of simple graphs. Recently it has been pointed out that restricting the attention to pairwise interactions is also a limitation, as the vast majority of complex systems include higher-order interactions that strongly affect their dynamics. Here, we propose hyper-diffusion on multiplex networks, a dynamical process in which diffusion on each single layer is coupled with the diffusion in other layers thanks to the presence of higher-order interactions occurring when there exists link overlap. We show that hyper-diffusion on a duplex network can be described by the Hyper-Laplacian in which the strength of four-body interactions among every set of four replica nodes connected by a multilink $(1,1)$ can be tuned by a parameter $δ_{11}\ge 0$. The Hyper-Laplacian reduces to the standard lower Laplacian, capturing pairwise interactions at the two layers, when $δ_{11}=0$. By combining tools of spectral graph theory, applied topology and network science we provide a general understanding of hyper-diffusion on duplex networks when $δ_{11}>0$, including theoretical bounds on the Fiedler and the largest eigenvalue of Hyper-Laplacians and the asymptotic expansion of their spectrum for $δ_{11}\ll1$ and $δ_{11}\gg1$. Although hyper-diffusion on multiplex networks does not imply a direct ``transfer of mass" among the layers (i.e. the average state of replica nodes in each layer is a conserved quantity of the dynamics), we find that the dynamics of the two layers is coupled as the relaxation to the steady state becomes synchronous when higher-order interactions are taken into account and the Fiedler eigenvalue of the Hyper-Laplacian is not localized in a single layer of the duplex network.

physics.soc-ph↗

arXiv2022

Synchronization induced by directed higher-order interactions

Non-reciprocal interactions play a crucial role in many social and biological complex systems. While directionality has been thoroughly accounted for in networks with pairwise interactions, its effects in systems with higher-order interactions have not yet been explored as deserved. Here, we introduce the concept of M-directed hypergraphs, a general class of directed higher-order structures, which allow to investigate dynamical systems coupled through directed group interactions. As an application we study the synchronization of nonlinear oscillators on 1-directed hypergraphs, finding that directed higher-order interactions can destroy synchronization, but also stabilize otherwise unstable synchronized states.

nlin.AO↗

arXiv2022

MultiSAGE: a multiplex embedding algorithm for inter-layer link prediction

Research on graph representation learning has received great attention in recent years. However, most of the studies so far have focused on the embedding of single-layer graphs. The few studies dealing with the problem of representation learning of multilayer structures rely on the strong hypothesis that the inter-layer links are known, and this limits the range of possible applications. Here we propose MultiSAGE, a generalization of the GraphSAGE algorithm that allows to embed multiplex networks. We show that MultiSAGE is capable to reconstruct both the intra-layer and the inter-layer connectivity, outperforming GraphSAGE, which has been designed for simple graphs. Next, through a comprehensive experimental analysis, we shed light also on the performance of the embedding, both in simple and in multiplex networks, showing that either the density of the graph or the randomness of the links strongly influences the quality of the embedding.

cs.LG↗

arXiv2022

Social interactions affect discovery processes

Our network of acquaintances determines how we get exposed to ideas, products, or cultural artworks (books, music, movies, etc.). Though this principle is part of our common sense, little is known about the specific pathways through which our peers influence our discovery processes and our experience of the new. Here, we fill this gap by investigating a data set containing the whole listening histories of a large, socially connected sample of users from the online music platform \emph{Last.fm}. We demonstrate that users exhibit highly heterogeneous discovery rates of new songs and artists and that their social neighborhood significantly influences their behavior. More explorative users tend to interact with peers more prone to explore new content. We capture this phenomenology in a modeling scheme where users are represented by random walkers exploring a graph of songs or artists and interacting with each other through their social links. Even starting from a uniform population of agents (no natural differences among the individuals), our model predicts the emergence of strong heterogeneous exploration patterns, with users clustered according to their musical tastes and propensity to explore. We contend our approach can pave the way to a quantitative approach to collective discovery processes.

physics.soc-ph↗

arXiv2021

Characterizing Learning Dynamics of Deep Neural Networks via Complex Networks

In this paper, we interpret Deep Neural Networks with Complex Network Theory. Complex Network Theory (CNT) represents Deep Neural Networks (DNNs) as directed weighted graphs to study them as dynamical systems. We efficiently adapt CNT measures to examine the evolution of the learning process of DNNs with different initializations and architectures: we introduce metrics for nodes/neurons and layers, namely Nodes Strength and Layers Fluctuation. Our framework distills trends in the learning dynamics and separates low from high accurate networks. We characterize populations of neural networks (ensemble analysis) and single instances (individual analysis). We tackle standard problems of image recognition, for which we show that specific learning dynamics are indistinguishable when analysed through the solely Link-Weights analysis. Further, Nodes Strength and Layers Fluctuations make unprecedented behaviours emerge: accurate networks, when compared to under-trained models, show substantially divergent distributions with the greater extremity of deviations. On top of this study, we provide an efficient implementation of the CNT metrics for both Convolutional and Fully Connected Networks, to fasten the research in this direction.

cs.LG↗

arXiv2021

The physics of higher-order interactions in complex systems

Complex networks have become the main paradigm for modelling the dynamics of interacting systems. However, networks are intrinsically limited to describing pairwise interactions, whereas real-world systems are often characterized by higher-order interactions involving groups of three or more units. Higher-order structures, such as hypergraphs and simplicial complexes, are therefore a better tool to map the real organization of many social, biological and man-made systems. Here, we highlight recent evidence of collective behaviours induced by higher-order interactions, and we outline three key challenges for the physics of higher-order systems.

physics.soc-ph↗