Empirical Networks Research Articles

Day-to-day dynamic origin–destination flow estimation using connected vehicle trajectories and automatic vehicle identification data

Dynamic vehicular origin–destination (OD) flow is a fundamental component of traffic network modeling and its estimation has long been studied. Although ideal observing conditions and behavioral assumptions are often indispensable for estimation, day-to-day traffic recurrences and variations are seldom utilized to improve the estimation performance. In this paper, we propose a new method to recover day-to-day dynamic OD flows using both connected vehicle (CV) trajectories and automatic vehicle identification (AVI) observations. The method involves two modules: the first module provides reliable prior OD flows given limited observations, while the second module seeks the optimal estimates based on the prior OD flows. In the first module, linear projection is extended to consider temporal and spatial variation of the CV penetration rate, and non-negative Tucker decomposition (NTD) is adopted to address the data sparsity issue caused by the low CV penetration rate. In the second module, a self-supervised learning model called the latency-constrained autoencoder (LCAE) is established to search for the optimal OD flows according to the priors with given robust latent features. To avoid local minima and ensure consistency between estimates, a novel algorithm called adaptive sub-sample correction (ASC) is proposed and integrated into the optimization process of LCAE, which can iteratively correct the most inconsistent samples based on the day-to-day traffic flow characteristics. The proposed method is examined on an empirical urban arterial network, a calibrated simulation network, and a synthetic large-scale grid network. Our results indicated that the proposed method requires very few AVI detectors and CV trajectories to achieve competitive estimation performance against two benchmark models. Furthermore, general robustness to several factors with respect to observing conditions and data quality was investigated, and satisfactory scalability was also demonstrated in terms of both estimation accuracy and computational cost.

Megabenthic standing stocks and organic carbon demand in a warming Arctic

Benthic megafauna (organisms large enough to be visible on seabed photographs) are regarded as important for carbon cycling in benthic habitats. They are a food source for many predators like fish and marine mammals and may stimulate carbon mineralization in sediment by bioturbation. However, few studies address these basic characteristics of megabenthos quantitatively. This study quantifies the spatial variability in standing stock (biomass) and functioning (secondary production, respiration and carbon demand) of benthic megafauna in fjords and on the continental shelf of Svalbard. Organisms were measured from sea bottom images to assess their biomass using length-weight relationships and volumetric methods, then respiration and production were estimated with empirical artificial neural network models. Significantly higher standing stock, secondary production, respiration, and carbon demand were found in fjords categorized as ‘cold’ (as defined by water temperature, prevailing water masses and ice-cover) than in the ‘warm’ ones. Cold fjords were dominated by Echinodermata, while in warm fjords Crustacea prevailed. All megafaunal community parameters were negatively correlated with bottom temperature. It was not possible to assess specific direct impacts of temperature, and indirect effects may be more relevant to our findings. These include temperature-driven changes in primary production, ice cover and ice-algae production or predation pressure from carnivores expanding their ranges northward. The progression of climate warming may affect megafaunal communities by reducing their biomass, production, and carbon demand and have profound effects on ecosystem functioning.

Effect of remote signal propagation in an empirical brain network.

Increasing evidence has shown that brain functions are seriously influenced by the heterogeneous structure of a brain network, but little attention has been paid to the aspect of signal propagation. We here study how a signal is propagated from a source node to other nodes on an empirical brain network by a model of bistable oscillators. We find that the unique structure of the brain network favors signal propagation in contrast to other heterogeneous networks and homogeneous random networks. Surprisingly, we find an effect of remote propagation where a signal is not successfully propagated to the neighbors of the source node but to its neighbors' neighbors. To reveal its underlying mechanism, we simplify the heterogeneous brain network into a heterogeneous chain model and find that the accumulation of weak signals from multiple channels makes a strong input signal to the next node, resulting in remote propagation. Furthermore, a theoretical analysis is presented to explain these findings.

Reinterpreting the relationship between number of species and number of links connects community structure and stability.

For 50 years, ecologists have examined how the number of interactions (links) scales with the number of species in ecological networks. Here, we show that the way the number of links varies when species are sequentially removed from a community is fully defined by a single parameter identifiable from empirical data. We mathematically demonstrate that this parameter is network-specific and connects local stability and robustness, establishing a formal connection between community structure and two prime stability concepts. Importantly, this connection highlights a local stability-robustness trade-off, which is stronger in mutualistic than in trophic networks. Analysis of 435 empirical networks confirmed these results. We finally show how our network-specific approach relates to the classical across-network approach found in literature. Taken together, our results elucidate one of the intricate relationships between network structure and stability in community networks.

Information Entropy Based on Propagation Feature of Node for Identifying the Influential Nodes

For understanding and controlling spreading in complex networks, identifying the most influential nodes, which can be applied to disease control, viral marketing, air traffic control, and many other fields, is of great importance. By taking the effect of the spreading rate on information entropy into account, we proposed an improved information entropy (IIE) method. Compared to the benchmark methods in the six different empirical networks, the IIE method has been found with a better performance on Kendall’s Tau and imprecision function under the Susceptible Infected Recovered (SIR) model. Especially in the Facebook network, Kendall’s Tau can grow by 120% as compared with the original IE method. And, there is also an equally good performance in the comparative analysis of imprecise functions. The imprecise functions’ value of the IIE method is smaller than the benchmark methods in six networks.

Building surrogate temporal network data from observed backbones.

Many systems of socioeconomic interests find a convenient representation in the form of temporal networks, i.e., sets of nodes and interactions occurring at specified times. In the corresponding data sets, however, crucial elements coexist with nonessential ones and noise. Several methods have thus been proposed to extract a "network backbone," i.e., the set of most important links in a network data set. The outcome of such methods can be seen as compressed versions of the original data. However, the question of how to practically use such reduced views of the data has not been tackled: for instance, using them directly in numerical simulations of processes on networks might lead to important biases. Overall, such reduced views of the data might not be actionable without an adequate decompression method. Here, we address this issue by putting forward and exploring several systematic procedures to build surrogate data from various kinds of temporal network backbones. In particular, we explore how much information about the original data needs to be retained alongside the backbone so that the surrogate data can be used in data-driven numerical simulations of spreading processes on a wide range of spreading parameters. We illustrate our results using empirical temporal networks with a broad variety of structures and properties. Our results give hints on how to best summarize complex data sets so that they remain actionable. Moreover, they show how ensembles of surrogate data with similar properties can be obtained from an original single data set, without any modeling assumptions.

Identification of multiple influential spreaders on networks by percolation under the SIR model.

Identification of multiple influential spreaders on complex networks is of great significance, which can help us speed up information diffusion and prevent disease from spreading to some extent. The traditional top-k strategy to solve an influence maximization problem based on node centrality is unsuitable for selecting several spreaders simultaneously because of influence overlapping. Besides, other heuristic methods have a poor ability to keep the balance between efficiency and computing time. In this paper, an efficient method is proposed to identify the decentralized influential spreaders on networks by edge percolation under the Susceptible-Infected-Recovered (SIR) model. Thanks to the average size of the connected component where one node is located under the edge percolation equivalent to the final spread range of this node under the SIR model approximately, it inspires us to choose suitable spreaders maximize the spread of influence. The experimental results show that our method has high efficiency compared with other benchmark methods on three synthetic networks and six empirical networks, and it also requires less time and cost.

Predicting partially observed processes on temporal networks by Dynamics-Aware Node Embeddings (DyANE)

Low-dimensional vector representations of network nodes have proven successful to feed graph data to machine learning algorithms and to improve performance across diverse tasks. Most of the embedding techniques, however, have been developed with the goal of achieving dense, low-dimensional encoding of network structure and patterns. Here, we present a node embedding technique aimed at providing low-dimensional feature vectors that are informative of dynamical processes occurring over temporal networks – rather than of the network structure itself – with the goal of enabling prediction tasks related to the evolution and outcome of these processes. We achieve this by using a lossless modified supra-adjacency representation of temporal networks and building on standard embedding techniques for static graphs based on random walks. We show that the resulting embedding vectors are useful for prediction tasks related to paradigmatic dynamical processes, namely epidemic spreading over empirical temporal networks. In particular, we illustrate the performance of our approach for the prediction of nodes’ epidemic states in single instances of a spreading process. We show how framing this task as a supervised multi-label classification task on the embedding vectors allows us to estimate the temporal evolution of the entire system from a partial sampling of nodes at random times, with potential impact for nowcasting infectious disease dynamics.

Parameter identification of complex network dynamics

Here, we present a novel approach called the “parameter identification of complex network dynamics” algorithm which combines elements of the sparse identification of nonlinear dynamics algorithm with a genetic algorithm to automatically and efficiently discover the underlying dynamics of complex networks from data with minimal domain-specific knowledge requirements. Testing the proposed algorithm on empirical complex network data verifies the accuracy and efficiency of this method compared to a purely evolutionary approach.

Mitigating COVID-19 Transmission in Schools With Digital Contact Tracing.

Precision mitigation of COVID-19 is in pressing need for postpandemic time with the absence of pharmaceutical interventions. In this study, the effectiveness and cost of digital contact tracing (DCT) technology-based on-campus mitigation strategy are studied through epidemic simulations using high-resolution empirical contact networks of teachers and students. Compared with traditional class, grade, and school closure strategies, the DCT-based strategy offers a practical yet much more efficient way of mitigating COVID-19 spreading in the crowded campus. Specifically, the strategy based on DCT can achieve the same level of disease control as rigid school suspensions but with significantly fewer students quarantined. We further explore the necessary conditions to ensure the effectiveness of DCT-based strategy and auxiliary strategies to enhance mitigation effectiveness and make the following recommendation: social distancing should be implemented along with DCT, the adoption rate of DCT devices should be assured, and swift virus tests should be carried out to discover asymptomatic infections and stop their subsequent transmissions. We also argue that primary schools have higher disease transmission risks than high schools and, thereby, should be alerted when considering reopenings.

Influence of Sound on Empirical Brain Networks

We analyze the influence of an external sound source in a network of FitzHugh–Nagumo oscillators with empirical structural connectivity measured in healthy human subjects. We report synchronization patterns, induced by the frequency of the sound source. We show that the level of synchrony can be enhanced by choosing the frequency of the sound source and its amplitude as control parameters for synchronization patterns. We discuss a minimum model elucidating the modalities of the influence of music on the human brain.

From foraging trails to transport networks: how the quality-distance trade-off shapes network structure.

Biological systems are typically dependent on transportation networks for the efficient distribution of resources and information. Revealing the decentralized mechanisms underlying the generative process of these networks is key in our global understanding of their functions and is of interest to design, manage and improve human transport systems. Ants are a particularly interesting taxon to address these issues because some species build multi-sink multi-source transport networks analogous to human ones. Here, by combining empirical field data and modelling at several scales of description, we show that pre-existing mechanisms of recruitment with positive feedback involved in foraging can account for the structure of complex ant transport networks. Specifically, we find that emergent group-level properties of these empirical networks, such as robustness, efficiency and cost, can arise from models built on simple individual-level behaviour addressing a quality-distance trade-off by the means of pheromone trails. Our work represents a first step in developing a theory for the generation of effective multi-source multi-sink transport networks based on combining exploration and positive reinforcement of best sources.

Cost-efficiency trade-offs of the human brain network revealed by a multiobjective evolutionary algorithm

It is widely believed that the formation of brain network architecture is under the pressure of optimal trade-off between reducing wiring cost and promoting communication efficiency. However, the questions of whether this trade-off exists in empirical human brain structural networks and, if so, how it takes effect are still not well understood. Here, we employed a multiobjective evolutionary algorithm to directly and quantitatively explore the cost-efficiency trade-off in human brain structural networks. Using this algorithm, we generated a population of synthetic networks with optimal but diverse cost-efficiency trade-offs. It was found that these synthetic networks could not only reproduce a large portion of connections in the empirical brain structural networks but also embed a resembling small-world organization. Moreover, the synthetic and empirical brain networks were found similar in terms of the spatial arrangement of hub regions and the modular structure, which are two important topological features widely assumed to be outcomes of cost-efficiency trade-offs. The synthetic networks had high robustness against random attacks as the empirical brain networks did. Additionally, we also revealed some differences between the synthetic networks and the empirical brain networks, including lower segregated processing capacity and weaker robustness against targeted attacks in the synthetic networks. These findings provide direct and quantitative evidence that the structure of human brain networks is indeed largely influenced by optimal cost-efficiency trade-offs. We also suggest that some additional factors (e.g., segregated processing capacity) might jointly determine the network organization with cost and efficiency.

Mathematical synthesis of the cortical circulation for the whole mouse brain—part II: Microcirculatory closure

Recent advancements in multiphoton imaging and vascular reconstruction algorithms have increased the amount of data on cerebrovascular circulation for statistical analysis and hemodynamic simulations. Experimental observations offer fundamental insights into capillary network topology but mainly within a narrow field-of-view typically spanning a small fraction of the cortical surface (less than 2%). In contrast, larger-resolution imaging modalities, like computed tomography (CT) or magnetic resonance imaging (MRI), have whole-brain coverage but capture only larger blood vessels, overlooking the microscopic capillary bed. To integrate data acquired at multiple length scales with different neuroimaging modalities and to reconcile brain-wide macroscale information with microscale multiphoton data, we developed a method for synthesizing hemodynamically equivalent vascular networks for the entire cerebral circulation. This computational approach is intended to aide in the quantification of patterns of cerebral blood flow and metabolism for the entire brain. In part I, we described the mathematical framework for image-guided generation of synthetic vascular networks covering the large cerebral arteries from the Circle of Willis through the pial surface network leading back to the venous sinuses. Here in part II, we introduce novel procedures for creating microcirculatory closure that mimics a realistic capillary bed. We demonstrate our capability to synthesize synthetic vascular networks whose morphometrics match empirical network graphs from three independent state-of-the art imaging labs using different image acquisition and reconstruction protocols. We also successfully synthesized twelve vascular networks of a complete mouse brain hemisphere suitable for performing whole brain blood flow simulations. Synthetic arterial and venous networks with microvascular closure allow whole brain hemodynamic predictions. Simulations across all length scales will potentially illuminate organ-wide supply and metabolic functions that are inaccessible to models reconstructed from image data with limited spatial coverage.

Combining multi-mode representations and ResNet for SAR target recognition

ABSTRACT This letter combines the multi-mode representations extracted by bidimensional empirical mode decomposition (BEMD) and deep residual networks (ResNet) for synthetic aperture radar (SAR) target recognition. Bidimensional intrinsic mode functions (BIMFs) are generated by BEMD to describe the target characteristics in SAR images. A special ResNet is designed and trained for each layer of BIMFs. The decisions from different BIMFs are linearly fused using a random weight matrix. Typical test scenarios are designed on the Moving and Stationary Target Acquisition and Recognition (MSTAR) dataset to examine the proposed method. The results validate the validity and robustness of the method.

Percolation transitions in growing networks under achlioptas processes: Analytic solutions

Networks are ubiquitous in diverse real-world systems. Many empirical networks grow as the number of nodes increases with time. Percolation transitions in growing random networks can be of infinite order. However, when the growth of large clusters is suppressed under some effects, e.g., the Achlioptas process, the transition type changes to the second order. However, analytical results for the critical behavior, such as the transition point, critical exponents, and scaling relations are rare. Here, we derived them explicitly as a function of a control parameter $m$ representing the suppression strength using the scaling ansatz. We then confirmed the results by solving the rate equation and performing numerical simulations. Our results clearly show that the transition point approaches unity and the order-parameter exponent $\beta$ approaches zero algebraically as $m \to \infty$, whereas they approach these values exponentially for a static network. Moreover, the upper critical dimension becomes $d_u=4$ for growing networks, whereas it is $d_u=2$ for static ones.

Assembly processes and co-occurrence relationships in the bacterioplankton communities of a large river system

The incoming water area of the Central Line Project of the South-to-North Water Diversion comprises the upper reaches of the Han River in China. Bacterioplankton are important members of river ecosystems. However, the mechanisms (deterministic or stochastic processes) regulating planktonic bacterial community structure in the upper Han River are not well understood. Herein, we documented the spatiotemporal patterns, assembly processes, and co-occurrence relationships of the planktonic bacterial communities in the Han River and its tributaries. Seasonal variations in bacterioplankton communities were clearly observed and were more obvious than spatial differences. The planktonic bacterial community showed a significant distance (dendritic distance) decay pattern during the two sampling periods, and the distance decay pattern was stronger in spring than in the other season. Network analysis showed that bacterial community networks displayed non-random co-occurrence patterns, and significant differences in the topological properties of empirical and random networks (average clustering coefficient, average path length, and average degree) were found. Compared to those of the tributaries, the network of the mainstem had the lowest modularity; this may be caused by the significant difference in biodiversity between the mainstem and the tributaries. Redundancy analysis revealed that pH exhibited significant effects on the planktonic bacterial community in both seasons. However, variation partitioning analysis showed that spatial factors (stochastic processes) contributed more to explaining community assembly than environmental factors during the two seasons. In addition, the neutral community model can also explain a large part of the community variation observed during the autumn and spring (R2 = 0.817 and 0.808, respectively). This study shows that seasonal variation in the bacterial community is greatly influenced by seasonal environmental variation. However, stochastic processes dominantly affect the structure of the bacterioplankton community within seasons. In order to effectively manage and protect microbial diversity, managers consider not only local species diversity, but also the effect of regional species dispersal.

The impact of noise and topology on opinion dynamics in social networks.

We investigate the impact of noise and topology on opinion diversity in social networks. We do so by extending well-established models of opinion dynamics to a stochastic setting where agents are subject both to assimilative forces by their local social interactions, as well as to idiosyncratic factors preventing their population from reaching consensus. We model the latter to account for both scenarios where noise is entirely exogenous to peer influence and cases where it is instead endogenous, arising from the agents’ desire to maintain some uniqueness in their opinions. We derive a general analytical expression for opinion diversity, which holds for any network and depends on the network’s topology through its spectral properties alone. Using this expression, we find that opinion diversity decreases as communities and clusters are broken down. We test our predictions against data describing empirical influence networks between major news outlets and find that incorporating our measure in linear models for the sentiment expressed by such sources on a variety of topics yields a notable improvement in terms of explanatory power.

A Resource Allocation Algorithm for Ultra-Dense Networks Based on Deep Reinforcement Learning

The resource optimization of ultra-dense networks (UDNs) is critical to meet the huge demand of users for wireless data traffic. But the mainstream optimization algorithms have many problems, such as the poor optimization effect, and high computing load. This paper puts forward a wireless resource allocation algorithm based on deep reinforcement learning (DRL), which aims to maximize the total throughput of the entire network and transform the resource allocation problem into a deep Q-learning process. To effectively allocate resources in UDNs, the DRL algorithm was introduced to improve the allocation efficiency of wireless resources; the authors adopted the resource allocation strategy of the deep Q-network (DQN), and employed empirical repetition and target network to overcome the instability and divergence of the results caused by the previous network state, and to solve the overestimation of the Q value. Simulation results show that the proposed algorithm can maximize the total throughput of the network, while making the network more energy-efficient and stable. Thus, it is very meaningful to introduce the DRL to the research of UDN resource allocation.

Digital proximity tracing on empirical contact networks for pandemic control

Digital contact tracing is a relevant tool to control infectious disease outbreaks, including the COVID-19 epidemic. Early work evaluating digital contact tracing omitted important features and heterogeneities of real-world contact patterns influencing contagion dynamics. We fill this gap with a modeling framework informed by empirical high-resolution contact data to analyze the impact of digital contact tracing in the COVID-19 pandemic. We investigate how well contact tracing apps, coupled with the quarantine of identified contacts, can mitigate the spread in real environments. We find that restrictive policies are more effective in containing the epidemic but come at the cost of unnecessary large-scale quarantines. Policy evaluation through their efficiency and cost results in optimized solutions which only consider contacts longer than 15–20 minutes and closer than 2–3 meters to be at risk. Our results show that isolation and tracing can help control re-emerging outbreaks when some conditions are met: (i) a reduction of the reproductive number through masks and physical distance; (ii) a low-delay isolation of infected individuals; (iii) a high compliance. Finally, we observe the inefficacy of a less privacy-preserving tracing involving second order contacts. Our results may inform digital contact tracing efforts currently being implemented across several countries worldwide.