Estimating the state of epidemics spreading with graph neural networks

Tomy, Abhishek; Razzanelli, Matteo; Lauro, Francesco Di; Rus, Daniela; Santina, Cosimo Della

doi:10.1007/s11071-021-07160-1

Cited by 16 publications

(12 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Recently, machine learning approaches have gained popularity. Tomy et al [40] exploit Graph Neural Networks (GNN) to solve the problem of inferring the state of the entire population by observing just a few individuals. GNN have also been used to forecast pandemic evolution [41]- [43],…”

Section: Models Of Epidemicsmentioning

confidence: 99%

Models for Digitally Contact-Traced Epidemics

2022

View full text Add to dashboard Cite

Contacts between people are the main drivers of contagious respiratory infections. For this reason, limiting and tracking contacts is a key strategy for controlling the COVID-19 epidemic. Digital contact tracing has been proposed as an automated solution to scale up traditional contact tracing. However, the required penetration of contact tracing apps within a population to achieve a desired target in controlling the epidemic is currently under discussion within the research community. In order to understand the effects of digital contact tracing, several mathematical models have been studied. In this article, we propose a novel compartmental SEIR model with which it is possible, differently from the models in the related literature, to derive closed-form conditions regarding the control of the epidemic. These conditions are a function of the penetration of contact tracing applications and testing efficiency. Closed-form conditions are crucial for the understandability of models, and thus for decision makers (including digital contact tracing designers) to correctly assess the dependencies within the epidemic. Feeding COVID-19 data to our model, we find that digital contact tracing alone can rarely tame the epidemic: for unrestrained COVID-19, this would require a testing turnaround of around 1 day and app uptake above 80% of the population, which are very difficult to achieve in practice. However, digital contact tracing can still be effective if complemented with other mitigation strategies, such as social distancing and mask-wearing.INDEX TERMS COVID-19, analytical models, digital contact tracing, testing efficiency, SEIR model, closed-form solutions, app uptake, epidemic controllability This article has been accepted for publication in IEEE Access.

show abstract

Section: Models Of Epidemicsmentioning

confidence: 99%

Models for Digitally Contact-Traced Epidemics

2022

View full text Add to dashboard Cite

show abstract

“…Além disso, nesse tipo de modelagem, a matriz de atributos dos vértices normalmente carrega o estado de saúde de cada indivíduo ou de grupos de pessoas (e.g., pessoas que vivem em uma dada região). Tomy et al (2022) utilizam dados de redes sociais para estimar a disseminação de Covid-19. Cada grafo não direcionado representa uma rede de usuários, onde cada indivíduo é associado a um vértice e o contato entre duas pessoas é representado por uma aresta.…”

Section: Previsão De Disseminação De Doençasunclassified

“…A i j = 1 se v i e v j estabelecem algum relacionamento ou contato e A i j = 0 caso contrário (Xiao et al, 2020), (Tomy et al, 2022), (Mondal et al, 2020), (Sanchez et al, 2021) e (Sanchez et al, 2021) Frequência A i j indica a frequência de transições/viagens entre v i e v j (Capanema et al, 2021b) e (La Gatta et al, 2020) Similaridade A i j representa o grau de similaridade entre as representações de v i e v j (Shen et al, 2018) Vizinhança A i j = 1 se v i e v j forem vizinhos e A i j = 0 caso contrário. Esse é um caso particular de relacionamento/contato para entidades geográficas (S. Hu et al, 2021) e (S. [G (1) ,G (2) , .…”

Section: Relacionamento/ Contatounclassified

Redes Neurais de Grafos no Contexto das Cidades Inteligentes

Capanema¹,

Silva²,

Loureiro³

2022

Minicursos Do XL Simpósio Brasileiro De Redes De Computadores E Sistemas Distribuídos

View full text Add to dashboard Cite

Graph neural networks, also known as GNNs, have been applied to solve problems in different domains, such as biology, chemistry, physics, natural language processing, computer vision, economics, among others. In particular, this class of neural network has been shown to be effective in modeling problems related to smart cities, such as traffic prediction; urban function classification of elements such as points of interest, roads and regions; forecasting the spread of diseases; autonomous agents; anomaly detection, among other activities. The objective of this chapter is to present the theoretical and practical foundations of graph neural networks in the context of smart cities. For this, the following topics are presented: types of tasks performed, taxonomy of GNNs layers, concepts, commonly used architectures, problems related to smart cities, modeling for the transformation of raw data to the graph structure and GNN structure, in addition to the implementation a model in practice. ResumoRedes neurais de grafos, também conhecidas como Graph Neural Networks (GNNs), têm sido aplicadas para resolver problemas em diferentes domínios, como biologia, química, física, processamento de linguagem natural, visão computacional, economia, dentre outros. Em particular, essa classe de rede neural tem-se mostrado eficaz na modelagem de problemas relacionados às cidades inteligentes, como predição de tráfego; classificação de função urbana de elementos como pontos de interesse, estradas e regiões; previsão de disseminação de doenças; agentes autônomos; detecção de anomalia, dentre outras atividades. O objetivo deste capítulo é apresentar os fundamentos teóricos e práticos das redes neurais de grafos no contexto das cidades inteligentes. Para isto, são apresentados os seguintes tópicos: tipos de tarefas realizadas, taxonomia das camadas GNNs, conceitos, arquiteturas comumente utilizadas, problemas relacionados

show abstract

“…Graph neural networks (GNNs) [4] have been one of the most popular tools to deal with this kind of data and can learn powerful representation from graph-structural data. GNNs can be used in many tasks including node classification [5], link prediction [6], graph classification [7] and recommendation systems [8] and many others [43]- [46].…”

Section: Introductionmentioning

confidence: 99%

Schatten Graph Neural Networks

Liu

Chen

et al. 2022

IEEE Access

View full text Add to dashboard Cite

Graph Neural Networks (GNNs) have been intensively studied in recent years because of their promising performance over graph-structural data and have provided assistance in many fields. Recalling recent works on graph neural networks, we found that imposing graph smoothing via Frobenius norm was proven to be effective in the architecture of graph neural networks from the standpoint of the graph signal processing. In this paper, we aim to model the graph smoothness of graph neural networks using a Schatten p-norm with p in the interval [1, 2) to characterize smoothness and propose a novel architecture called Schatten graph neural networks. This architecture stems from a primal-dual solution scheme for a graph signal denoising problem. There is difficulty in solving subproblems with respect to the Schatten p-norm. We propose a fixed point iteration scheme and prove that it tracks with the linear convergence rate with solid mathematical analysis. Extensive experiments demonstrate the effectiveness of the proposed architecture of graph neural networks and their robustness to the graph adversarial attacks.

show abstract

Estimating the state of epidemics spreading with graph neural networks

Cited by 16 publications

References 53 publications

Models for Digitally Contact-Traced Epidemics

Models for Digitally Contact-Traced Epidemics

Redes Neurais de Grafos no Contexto das Cidades Inteligentes

Schatten Graph Neural Networks

Contact Info

Product

Resources

About