arXiv subjects

Vito Latora

Publications and source records attributed to Vito Latora.

At least 19 recordsLinked to original sources

arXiv2026

Dynamical processes and emergent behaviors in multiplex networks

Over the last two decades, network science has greatly advanced our understanding of how the collective behaviors of a complex system emerge from the interactions among its basic units. Multiplex networks, i.e. networks with many layers, whose nodes are in one-to-one correspondence, provide a more realistic description for social, biological and ecological systems where multiple types of interactions coexist. After a brief introduction on how to model the architecture of multiplex networks, we present a complete overview of the different dynamics which can unfold over these structures. We present a unified framework to describe dynamical processes such as percolation, reaction-diffusion, synchronization, epidemic spreading, social dynamics and games on multiplex networks, as well as the coupled evolution of different dynamical processes, and the coevolution of a process with the network structure. Our focus is on truly-multiplex collective behaviors, i.e., all those phenomena which cannot emerge on the corresponding aggregated networks, or when the different layers of these systems are considered in isolation. We identify three main mechanisms leading to new collective behaviors: the existence of structural correlations across layers, the presence of dynamical correlations in the processes taking place at the different layers, and the dynamical interplay of inter- and intra-layer interactions. We conclude with a summary of the main takeaways from a decade of work in the field.

physics.soc-ph↗

arXiv2026

Optimal network structure for collective performance with strategic information sharing

Information sharing between individuals is crucial to improve performance in collective tasks. However, in a competitive world, individuals may be reluctant to share information with the others, and it is still unclear how the presence of strategic behaviors affects the collective performance of a group. In this study, we introduce an evolutionary game modeling the dynamics of individual behaviors in a collective estimation task. The individuals are organized in a network and have to guess the distribution of ball colors in a box. Each of them samples a given number of balls and can strategically decide whether to share or not this information with its neighbors. We develop a framework that allows to investigate analytically how the collective performance depends on the network structure. We find that the optimal network results from a trade-off between the sharing rate and the way the information is integrated in the network. We further reveal that there exists an intermediate average degree for each type of network maximizing the collective performance. In addition to the uniform case, we consider the case of non-homogeneous allocations of the number of individual samples, showing that the largest collective performance is obtained when the number of ball extracted by an individual is inversely proportional to its degree.

physics.soc-ph↗

arXiv2026

Predicting Dynamics of Ultra-Large Complex Systems by Inferring Governing Equations

Predicting the behavior of ultra-large complex systems, from climate to biological and technological networks, is a central unsolved challenge. Existing approaches face a fundamental trade-off: equation discovery methods provide interpretability but fail to scale, while neural networks scale but operate as black boxes and often lose reliability over long times. Here, we introduce the Sparse Identification Graph Neural Network, a framework that overcome this divide by allowing to infer the governing equations of large networked systems from data. By defining symbolic discovery as edge-level information, SIGN decouples the scalability of sparse identification from network size, enabling efficient equation discovery even in large systems. SIGN allows to study networks with over 100,000 nodes while remaining robust to noise, sparse sampling, and missing data. Across diverse benchmark systems, including coupled chaotic oscillators, neural dynamics, and epidemic spreading, it recovers governing equations with high precision and sustains accurate long-term predictions. Applied to a data set of time series of temperature measurements in 71,987 sea surface positions, SIGN identifies a compact predictive network model and captures large-scale sea surface temperature conditions up to two years in advance. By enabling equation discovery at previously inaccessible scales, SIGN opens a path toward interpretable and reliable prediction of real-world complex systems.

cs.LG↗

arXiv2026

Group dynamics shape contagion onsets and multistable active phases under collective reinforcement

Group-based reinforcement can induce discontinuous transitions from inactive to active phases in higher-order contagion models. However, these results are typically obtained on static interaction structures or within mean-field approximations that neglect temporal changes in group composition. Here, we show that group dynamics is not a secondary effect but a central aspect that determines the macroscopic transition class of higher-order contagion processes. We develop an analytically tractable approximate master equation model that effectively interpolates between quenched and mean-field limits through a group composition turnover rate. Our results reveal the rich impact of time-varying structures: it can induce discontinuous phase transition, broaden the bistable region, and at the same time promote or suppress contagion near criticality. Moreover, when real-world turnover rates and group-size heterogeneity are taken into account, the system exhibits a qualitatively richer phase diagram with four distinct dynamical phases, combining continuous or discontinuous transitions with localized or delocalized activity. In localized regimes, we uncover multistable active phases with multiple coexisting active states, which are observed in neither the annealed nor the quenched limits, and extend classical absorbing-active bistability. Finally, we demonstrate that the emergence of discontinuous transitions in real-world systems requires stronger nonlinear reinforcement than previously thought, indicating that simulations in static structures can yield qualitatively misleading predictions.

physics.soc-ph↗

arXiv2026

Systematic Characterization of Minimal Deep Learning Architectures: A Unified Analysis of Convergence, Pruning, and Quantization

Deep learning networks excel at classification, yet identifying minimal architectures that reliably solve a task remains challenging. We present a computational methodology for systematically exploring and analyzing the relationships among convergence, pruning, and quantization. The workflow first performs a structured design sweep across a large set of architectures, then evaluates convergence behavior, pruning sensitivity, and quantization robustness on representative models. Focusing on well-known image classification of increasing complexity, and across Deep Neural Networks, Convolutional Neural Networks, and Vision Transformers, our initial results show that, despite architectural diversity, performance is largely invariant and learning dynamics consistently exhibit three regimes: unstable, learning, and overfitting. We further characterize the minimal learnable parameters required for stable learning, uncover distinct convergence and pruning phases, and quantify the effect of reduced numeric precision on trainable parameters. Aligning with intuition, the results confirm that deeper architectures are more resilient to pruning than shallower ones, with parameter redundancy as high as 60%, and quantization impacts models with fewer learnable parameters more severely and has a larger effect on harder image datasets. These findings provide actionable guidance for selecting compact, stable models under pruning and low-precision constraints in image classification.

cs.LG↗

arXiv2025

Endogenous Feedback in Coevolutionary Games Reshapes the Stability of Cooperation

In Evolutionary game theory the payoffs are typically fixed or shaped by external environmental variables. Here, we introduce an endogenous-feedback model in which the game played coevolves directly with the population state: the payoff matrix is a time-dependent function of the level of cooperation. This allows strategic incentives to be continuously modified by the collective behavior they generate. Even in the simplest case of linear and instantaneous feedback, the model reveals feedback-induced regimes, termed chimera games, in which stable cooperation arises despite being incompatible with the predictions of standard fixed-game dynamics. We further show that delayed feedback can destabilize these equilibria and generate sustained oscillations, while nonlinear feedback reshapes equilibrium structure and introduces path dependence. Our results show how cooperation can be promoted, suppressed, or destabilized by incentives generated endogenously by the very same population's collective behavior. We conclude by outlining how our framework connects to real-world systems shaped by endogenous feedback.

physics.soc-ph↗

arXiv2025

Symbolic Higher-Order Analysis of Multivariate Time Series

Identifying patterns of relations among the units of a complex system from measurements of their activities in time is a fundamental problem with many practical applications. Here, we introduce a method that detects dependencies of any order in multivariate time series data. The method first transforms a multivariate time series into a symbolic sequence, and then extract statistically significant strings of symbols through a Bayesian approach. Such motifs are finally modelled as the hyperedges of a hypergraph, allowing us to use network theory to study higher-order interactions in the original data. When applied to neural and social systems, our method reveals meaningful higher-order dependencies, highlighting their importance in both brain function and social behaviour.

physics.soc-ph↗

arXiv2025

Drivers of cooperation in social dilemmas on higher-order networks

Understanding cooperation in social dilemmas requires models that capture the complexity of real-world interactions. While network frameworks have provided valuable insights to model the evolution of cooperation, they are unable to encode group interactions properly. Here, we introduce a general higher-order network framework for multi-player games on structured populations. Our model considers multi-dimensional strategies, based on the observation that social behaviours are affected by the size of the group interaction. We investigate dynamical and structural coupling between different orders of interactions, revealing the crucial role of nested multilevel interactions, and showing how such features can enhance cooperation beyond the limit of traditional models with uni-dimensional strategies. Our work identifies the key drivers promoting cooperative behaviour commonly observed in real-world group social dilemmas.

physics.soc-ph↗

arXiv2025

Hyperedge Overlap drives Synchronizability of Systems with Higher-Order interactions

The microscopic organization of dynamical systems coupled via higher-order interactions plays a pivotal role in understanding their collective behavior. In this paper, we introduce a framework for systematically investigating the impact of the interaction structure on dynamical processes. Specifically, we develop an hyperedge overlap matrix whose elements characterize the two main aspects of the microscopic organization of higher-order interactions: the inter-order hyperedge overlap (non-diagonal matrix elements) and the intra-order hyperedge overlap (encapsulated in the diagonal elements). This way, the first set of terms quantifies the extent of superposition of nodes among hyperedges of different orders, while the second focuses on the number of nodes in common between hyperedges of the same order. Our findings indicate that large values of both types of hyperedge overlap hinder synchronization stability, and that the larger is the order of interactions involved, the more important is their role. Our findings also indicate that the two types of overlap have qualitatively distinct effects on the dynamics of coupled chaotic oscillators. In particular, large values of intra-order hyperedge overlap hamper synchronization by favoring the presence of disconnected sets of hyperedges, while large values of inter-order hyperedge overlap hinder synchronization by increasing the number of shared nodes between groups converging on different trajectories, without necessarily causing disconnected sets of hyperedges.

physics.soc-ph↗

arXiv2024

The networks of ingredient combinations as culinary fingerprints of world cuisines

Investigating how different ingredients are combined in popular dishes is crucial to uncover the principles behind food preferences. Here, we use data from public food repositories and network analysis to characterize and compare worldwide cuisines. Ingredients are first grouped into broader types, and each cuisine is then represented as a network in which nodes correspond to ingredient types and weighted links describe how frequently pairs of types co-occur in recipes. Cuisines differ not only in the popularity of ingredient types and range of recipe sizes, but also in the structural organization of ingredient-type combinations. By analyzing these networks, we uncover distinctive patterns of type associations that serve as culinary fingerprints. For example, European cuisines typically distribute ingredients across different types, whereas certain Asian and South American traditions emphasize one dominant type, such as vegetables or spices. The essence of these patterns is well captured by the networks' maximum spanning trees, which offer a simplified yet representative backbone for each cuisine. We demonstrate that both these full and simplified network representations enable machine learning models to identify cuisines from subsets of recipes with very high accuracy. Networks of ingredient combinations also cluster global cuisines into meaningful geo-cultural groups, reflecting shared patterns in culinary traditions. More broadly, our study offers novel insights into the structure of world cuisines, enabling data-driven approaches to their characterization, cross-cultural comparison, and potential adaptation.

physics.soc-ph↗

arXiv2024

On Learnable Parameters of Optimal and Suboptimal Deep Learning Models

We scrutinize the structural and operational aspects of deep learning models, particularly focusing on the nuances of learnable parameters (weight) statistics, distribution, node interaction, and visualization. By establishing correlations between variance in weight patterns and overall network performance, we investigate the varying (optimal and suboptimal) performances of various deep-learning models. Our empirical analysis extends across widely recognized datasets such as MNIST, Fashion-MNIST, and CIFAR-10, and various deep learning models such as deep neural networks (DNNs), convolutional neural networks (CNNs), and vision transformer (ViT), enabling us to pinpoint characteristics of learnable parameters that correlate with successful networks. Through extensive experiments on the diverse architectures of deep learning models, we shed light on the critical factors that influence the functionality and efficiency of DNNs. Our findings reveal that successful networks, irrespective of datasets or models, are invariably similar to other successful networks in their converged weights statistics and distribution, while poor-performing networks vary in their weights. In addition, our research shows that the learnable parameters of widely varied deep learning models such as DNN, CNN, and ViT exhibit similar learning characteristics.

cs.LG↗

arXiv2024

Deep Neural Networks via Complex Network Theory: a Perspective

Deep Neural Networks (DNNs) can be represented as graphs whose links and vertices iteratively process data and solve tasks sub-optimally. Complex Network Theory (CNT), merging statistical physics with graph theory, provides a method for interpreting neural networks by analysing their weights and neuron structures. However, classic works adapt CNT metrics that only permit a topological analysis as they do not account for the effect of the input data. In addition, CNT metrics have been applied to a limited range of architectures, mainly including Fully Connected neural networks. In this work, we extend the existing CNT metrics with measures that sample from the DNNs' training distribution, shifting from a purely topological analysis to one that connects with the interpretability of deep learning. For the novel metrics, in addition to the existing ones, we provide a mathematical formalisation for Fully Connected, AutoEncoder, Convolutional and Recurrent neural networks, of which we vary the activation functions and the number of hidden layers. We show that these metrics differentiate DNNs based on the architecture, the number of hidden layers, and the activation function. Our contribution provides a method rooted in physics for interpreting DNNs that offers insights beyond the traditional input-output relationship and the CNT topological analysis.

cs.LG↗

arXiv2023

Evolutionary game selection creates cooperative environments

The emergence of collective cooperation in competitive environments is a well-known phenomenon in biology, economics, and social systems. While most evolutionary game models focus on the evolution of strategies for a fixed game, how strategic decisions coevolve with the environment has so far mostly been overlooked. Here, we consider a game selection model where not only the strategies but also the game can change over time following evolutionary principles. Our results show that coevolutionary dynamics of games and strategies can induce novel collective phenomena, fostering the emergence of cooperative environments. When the model is taken on structured populations the architecture of the interaction network can significantly amplify pro-social behavior, with a critical role played by network heterogeneity and the presence of clustered groups of similar players, distinctive features observed in real-world populations. By unveiling the link between the evolution of strategies and games for different structured populations, our model sheds new light on the origin of social dilemmas ubiquitously observed in real-world social systems.

physics.soc-ph↗

arXiv2023

The Concept of Decentralization Through Time and Disciplines: A Quantitative Exploration

Decentralization is a pervasive concept found across disciplines, including Economics, Political Science, and Computer Science, where it is used in distinct yet interrelated ways. Here, we develop and publicly release a general pipeline to investigate the scholarly history of the term, analysing 425,144 academic publications that refer to (de)centralization. We find that the fraction of papers on the topic has been exponentially increasing since the 1950s. In 2021, 1 author in 154 mentioned (de)centralization in the title or abstract of an article. Using both semantic information and citation patterns, we cluster papers in fields and characterize the knowledge flows between them. Our analysis reveals that the topic has independently emerged in the different fields, with small cross-disciplinary contamination. Moreover, we show how Blockchain has become the most influential field about 10 years ago, while Governance dominated before the 1990s. In summary, our findings provide a quantitative assessment of the evolution of a key yet elusive concept, which has undergone cycles of rise and fall within different fields. Our pipeline offers a powerful tool to analyze the evolution of any scholarly term in the academic literature, providing insights into the interplay between collective and independent discoveries in science.

physics.soc-ph↗

arXiv2023

The importance of quality in austere times: University competitiveness and grant income

After 2009 many governments implemented austerity measures, often restricting science funding. Did such restrictions further skew grant income towards elite scientists and universities? And did increased competition for funding undermine participation? UK science funding agencies significantly reduced numbers of grants and total grant funding in response to austerity, but surprisingly restrictions of science funding were relaxed after the 2015 general election. Exploiting this natural experiment, we show that conventional measures of university competitiveness are poor proxies for competitiveness. An alternative measure of university competitiveness, drawn from complexity science, captures the highly dynamical way in which universities engage in scientific subjects. Building on a data set of 43,430 UK funded grants between 2006 and 2020, we analyse rankings of UK universities and investigate the effect of research competitiveness on grant income. When austerity was relaxed in 2015 the elasticity of grant income w.r.t. research competitiveness fell, reflecting increased effort by researchers at less competitive universities. These scientists increased number and size of grant applications, increasing grant income. The study reveals how funding agencies, facing heterogeneous competitiveness in the population of scientists, affect research effort across the distribution of competitiveness.

econ.GN↗

arXiv2023

A pair-based approximation for simplicial contagion

Higher-order interactions play an important role in complex contagion processes. Mean-field approximations have been used to characterize the onset of spreading in the presence of group interactions. However, individual-based mean-field models are unable to capture correlations between different subsets of nodes, which can significantly influence the dynamics of a contagion process. In this paper, we introduce a pair-based mean-field approximation that allows to study the dynamics of a SIS model on simplicial complexes by taking into account correlations at the level of pairs of nodes. %by taking into account dynamical correlations emerging in groups of nodes. Compared to individual-based mean-field approaches, the proposed approximation yields more accurate predictions of the dynamics of contagion processes on simplicial complexes. Specifically, the pair-based mean-field approximation provides higher accuracy in predicting the extent of the region of bistability, the type of transition from disease-free to endemic state, and the average time evolution of the fraction of infected individuals. Crucially, the pair-based approximation correctly predicts that the onset of the epidemic outbreak in simplicial complexes depends on the strength of higher-order interactions. Overall, our findings highlight the importance of accounting for pair correlations when investigating contagion processes in the presence of higher-order interactions.

physics.soc-ph↗

arXiv2023

The dynamics of higher-order novelties

Studying how we explore the world in search of novelties is key to understand the mechanisms that can lead to new discoveries. Previous studies analyzed novelties in various exploration processes, defining them as the first appearance of an element. However, novelties can also be generated by combining what is already known. We hence define higher-order novelties as the first time two or more elements appear together, and we introduce higher-order Heaps' exponents as a way to characterize their pace of discovery. Through extensive analysis of real-world data, we find that processes with the same pace of discovery, as measured by the standard Heaps' exponent, can instead differ at higher orders. We then propose to model an exploration process as a random walk on a network in which the possible connections between elements evolve in time. The model reproduces the empirical properties of higher-order novelties, revealing how the network we explore changes over time along with the exploration process.

physics.soc-ph↗

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↗