key: cord-0779251-222tzmjj authors: Jayatilaka, G. C.; Hassan, J.; Marikkar, U.; Perera, R.; Sritharan, S.; Weligampola, H.; Ekanayake, M.; Godaliyadda, R.; Ekanayake, P.; Herath, V.; Godaliyadda, G. M. D.; Rathnayake, A.; Dharmaratne, S. D.; Ekanayake, J. title: Use of Artificial Intelligence on spatio-temporal data to generate insights during COVID-19 pandemic: A Review date: 2020-11-24 journal: nan DOI: 10.1101/2020.11.22.20232959 sha: 0bf24b0bce0150bb143b0b9ed9e9fdd076e4c611 doc_id: 779251 cord_uid: 222tzmjj The COVID-19 pandemic, within a short time span, has had a significant impact on every aspect of life in almost every country on the planet. As it evolved from a local epidemic isolated to certain regions of China, to the deadliest pandemic since the influenza outbreak of 1918, scientists all over the world have only amplified their efforts to combat it. In that battle, Artificial Intelligence, or AI, with its wide ranging capabilities and versatility, has played a vital role and thus has had a sizable impact. In this review, we present a comprehensive analysis of the use of AI techniques for spatio-temporal modeling and forecasting and impact modeling on diverse populations as it relates to COVID-19. Furthermore, we catalogue the articles in these areas based on spatio-temporal modeling, intrinsic parameters, extrinsic parameters, dynamic parameters and multivariate inputs (to ascertain the penetration of AI usage in each sub area). The manner in which AI is used and the associated techniques utilized vary for each body of work. Majority of articles use deep learning models, compartment models, stochastic methods and numerous statistical methods. We conclude by listing potential paths of research for which AI based techniques can be used for greater impact in tackling the pandemic. Temporal forecasting of COVID-19 infection cases have been evaluated using multiple deep learning models. A multi-step prediction using the neural network models -Long Short Term Memory (LSTM), Convolutional LSTM (ConvLSTM), Bidirectional LSTM (Bi-LSTM)-as a case study for Canada, 16 different states in India, 17 and USA 18 have been studied. The Bi-LSTM model and the ConvLSTM models are shown to outperform the other LSTM variants with a Mean Absolute Percentage Error (MAPE) of 3% for day ahead predictions. Accuracy of variants of LSTMs are further compared with Gated Recurrent Units (GRU), a Variational Auto-Encoder (VAE), and Support Vector Regression (SVR) models. 19, 20 The models are evaluated on multiple countries where substantial data is available (USA, Italy, Spain, China, Australia) to validate that the Bi-LSTM and VAE (MAPE less than 3%) models outperform the conventional models. A data driven forecasting method has been developed to estimate the number of positive cases of COVID-19 in India for the next 30 days. 21 The number of recovered cases, daily positive cases, and deceased have also been estimated using LSTMs and curve fitting. In addition, a machine learning based random forest model has been used to forecast the number of COVID-19 cases with a mean correlation coefficient R 2 of 0·914. 22 A novel statistical method has been developed to forecast the spread of COVID-19 globally. This method considers lead-lag effects between different time series using the dynamic time warping technique to analyze non linear relationships among nations. 23 Thus, it was able to determine underlying causal relationships such as the origin of the virus and the forerunner in given regions. A Non linear Autoregressive Artificial Neural Networks (NAR-ANN) and Auto Regressive Integrated Moving Average (ARIMA) model was used to forecast the spread of COVID-19 in Egypt. 24 An analysis of COVID-19 deaths using reduced space Gaussian process regression has shown a correlation coefficient of 98·9%. 25 In an attempt to solve the patient triage problem related to COVID-19, researchers in Italy have formulated a discrete time markov chain model to predict the number of COVID-19 cases, to thereby, efficiently allocate ICU resources across the country. 14 The critical need for COVID-19 forecasting comes from the need to understand how to implement containment strategies while balancing its impact on the country's economy and thereby the livelihood of people. From an economical impact standpoint, a modified ARIMA model was developed to forecast the number of COVID-19 cases and the stock market in Spain 26 for the given period. ARIMA has also been ensembled with Wavelet Transforms to model stationary and non-stationary trends using a 10-day forecast. 27 Epidemic modeling in the past has extensively used compartment models, especially the Susceptible -Infective -Removed (SIR) model and its variants. Researchers in China have implemented a dynamic Susceptible Exposed Infectious Removed (SEIR) model along with an Artificial Intelligence model to predict the COVID-19 spread in China and identify underlying patterns in the spread. 28 A modified SEIR model has been implemented to forecast the spread of COVID-19 and its burden on hospital care under varying social distancing conditions. 29 The study concludes that the best parameter to assess the effectiveness of 4 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 24, 2020. confinement and risk of virus diffusion, is the average number of daily contacts in a population (c). Researchers have focused on modeling and forecasting the spread of COVID-19 in the spatio-temporal domain to identify not only ''how much" the disease has spread but also ''where" it has spread to. A study conducted for Germany has developed a memory based integro-differential network model to predict spatio-temporal outbreak dynamics such as number of infections, hospitalization rates, and demands on ICU due to COVID-19 at state level. 13 The model takes into account effects of different containment strategies, event and contact restrictions, and different courses the infection may take; which is not possible when dealing with traditional SIR models. A novel approach is taken using Moran's I test to find regional correlations of COVID-19 cases in China. 30, 31 The spatial association between states are modelled into the following six types: 30 1) Sharing of borders 2) Euclidean (Shortest) distance 3) Population 4) Population density 5) Number of doctors and hospitals 6) Number of medical beds. Results show high positive correlation between an infected region and its adjacent regions for the first five models. This method is useful for predicting outbreaks in specific regions, which in turn will allow policymakers to take proactive measures to prevent COVID-19 spread to adjacent regions. The county level spread of COVID-19 in relation to the healthcare capacities in Ohio, USA is predicted and the results show that the disease spreads much faster in counties that facilitate air transportation to other counties. 32 This shows that similar to attributes such as population, underlying spatio-temporal attributes such as inter-state travel can be employed to predict the dynamic spread of the pandemic. In this section, we review methods that group the population into clusters based on different factors such as occupation, age and other such parameters and analyze COVID-19 spread in those demographics. An overview of what characterizes the population diversity is given in Fig. 4 . The importance of this diversity/context awareness of information for handling the pandemic across the globe has been studied. 33 Even within similar geographies, socio-economic factors have been shown to have an effect on the spread and severity of the pandemic. 34 This work shows why the analysis, prediction and mitigation of COVID-19 related issues should be done considering the context (diversity of the population) for best results. Models developed to handle diversity awareness are mostly mappings between higher dimensional spaces (in cases where the context is passed as an input) with the capability to capture the multi-modal behavioral of data. We see multi-variate statistics, time series models, artificial neural networks and other AI models being used in this landscape. Key challenges in this domain are handling non-uniformly sampled data, incomplete data and figuring out the context without it being readily fed into the models. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The parameters taken into account to characterize diversity are analyzed in the following subsections (as given in Table 1 ). A summary of this section is given in Table 2 . A large-scale analysis of the severity of COVID-19 among the population, given the existing statistics of the medical conditions as per the Global Burden of Disease and Population census among the world, is done using statistical analysis coupled with resampling techniques. 35 This study shows the importance of utilizing larger data sources to generate insights during the pandemic. Another study aiming to uncover a similar correlation between pre-existing conditions and COVID-19 mortality was conducted on individual patient data from Wuhan using linear regression and statistical techniques. 36 In this section, we discuss the research work attempting to uncover the differential impact of COVID-19 on different age groups. The age dependence of COVID-19 's impact on children is studied using statistical analysis to thereby generate predictions on the population as a whole by considering data from USA. 37 This work goes on to generate recommendations based on the predictions on how best to prepare for large scale outbreaks. The age stratified CFR rate (case fatality rate --the ratio of deaths to the number of total cases) is analyzed using basic statistics. 38 Another study performs an in-depth analysis on a larger dataset from around the world (China's statistics coupled with the rest of the world). 39 The study also sheds light into the age dependence of healthcare in an overwhelmed system. From the literature, we conclude that age dependence is mostly analyzed using classical algorithms, as opposed to more modern techniques. The spread of COVID-19 has been analyzed with regard to environmental factors such as temperature and humidity. One study uses a spatial Seemingly Unrelated Regressions (SUR) to model the COVID-19 spread as an inter-regional contagion process. 40 This doesn't show the effect of temperature and sunshine on the behavior of the virus, but instead analyzes the people's behavioral changes based on environmental factors, and thereby the spread of the virus as a result of these changes. Another similar study (using ANN based differential equations and particle swarm algorithms) shows that population density (followed by relative humidity) has a higher impact on the spread of COVID-19 in comparison to temperature and wind speed. 41 6 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 24, 2020. ; Statistical models are used to show the absence of a correlation between the daily temperature and case counts in Spain. 42 Similarly, the performance of ANN based COVID-19 prediction is evaluated to show how absence of weather data does not make a difference. 43 This section discusses the effect of social, economic and political dependencies on the spread and severity of COVID-19. Moran's I test is widely used to analyze spatial correlations in pandemics. One such study computes the correlation between testing and COVID-19 positive rates, and the socio-economic status in zip code level clusters of New York. The study recommends healthcare resources to be directed towards areas where low testing and high positive rates are prevalent. 44 Another study explores spatial relationships between socio-demographics and the spread of COVID-19 in three regions of Texas using ordinary least squares (OLS) and geographically weighted regression (GWR). The authors quantify the influence of poverty on COVID-19 infection density for a given region. In turn, it helps policy-makers direct healthcare resources and provide more emphasis towards low-income regions in a state. 45 The effect of the percentage of African-American population on COVID-19 disease and death rates, for a given county in the United States is analyzed. Results indicated that nearly 90% of counties with a disproportionately higher number of African-Americans had a higher percentage of cases and deaths. The relationship between this socio-demographic and COVID-19 was fit to a Negative Binomial Regression (NBR) model, which can forecast the number of future cases in relation to income and racial parameters in a given county. 46 A case study of South Korea evaluates lockdown policies to find noteworthy COVID-19 clusters in space and time. It was observed that clusters were contained faster as time progressed (i.e. earlier outbreaks took time to contain but clusters that started later were contained quickly). This was evidence of how government policies can affect the spread of COVID-19. The model also took into account the population density of districts (higher population density implies higher disease spread). 47 Similarly, time, space and population demographics are used to model the COVID-19 outbreak containment in Italy. 48 A prediction modeling study has been conducted for 16 countries from diverse geographical regions, to measure the effect of Non-Pharmaceutical-Interventions (NPI) on the spread of COVID-19. 49 The authors analyze the effects of no intervention, cyclic mitigation or suppression measures followed by relaxation period. The effects on ICU admissions and death is evaluated for different measures. A similar study has been conducted for North America, Sweden 50 and China, 51 as they have taken contrasting policies to battle COVID-19. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 24, 2020. ; https://doi.org/10.1101/2020.11.22.20232959 doi: medRxiv preprint In this section we focus on travelling and mass population flow between cities during the COVID-19 pandemic prior to the lockdown measures and its impact. A recent study found correlations between the number of persons migrating from Wuhan to other cities and the number of cases in those cities. Using this, a risk analysis was developed using a Bayesian space-time model to quantify the risk of a person migrating from Wuhan spreading COVID-19 to another population. 52 While the similar scenario in India's mass migration has been studied, 53 AI is yet to be used in the Indian situation. Dynamic Parameters have also been used to analyze the spread of COVID-19. 54 This study identifies the regions with high risk of early transmission, and forecasts the distribution of COVID-19 cases from Wuhan, using statistical methods on population flow data. Modeling the effect of population flow with consideration on to the geography has outperformed the naive population flow models in terms of accuracy. The former was done by a neural network that can absorb time series information of the patient counts and travel statistics, as well as knowledge about the geographical hierarchy of countries and continents. 55 A study discusses the impact of travel restrictions on the spread of COVID-19 on a regionally connected population. To identify the spread of the disease, a Bayesian approach is used. The results show that travel restrictions had a minimal effect unless they were supplemented with other behavioral changes to mitigate the disease spread. 56 The methods in these articles attempt to model the complex relationships between different parameters and uncover correlations that might not be obvious in traditional statistical analysis. These models are designed based on both human intuition and AI. To identify the socio-economic variables most influential towards COVID-19 fatalities, a spatial regression approach is used. 57 Out of 28 possible demographics, population density, income, and poverty rates were found to have the highest correlation with infection rates and fatalities in a given region. This study also models fatalities of a given region using these reduced socio-economic variables. The reduction of variables vastly helps real-time forecasting due to simplicity and reduced computational cost. Another study developed a real-time risk assessment method for future cases in terms of fatality rate of 8 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 24, 2020. an affected country, using population density, healthcare resources, and government enforced preventive measures. The number of future cases was predicted using an ARIMA model, and a regression tree algorithm was used to validate the argument that the aforementioned variables have the highest impact on fatality rates in those countries. 58 Using the information of COVID-19 spread in the city of Lombardy, an individual level model for transmission that depends on geographical location, age, household structure, and current medical conditions The literature survey has uncovered two key application areas in which AI is used--predicting the future impact of COVID-19 on populations and uncovering the differential impact of COVID-19 on diverse segments of the population. The performance of predictive algorithms have been limited by a range of factors (Section 3.1). Impact modeling has generated interesting insights on the virus as well as the society at large (Section 3.2). Despite the recent growth of AI, many existing solutions rely on traditional methodologies, without any automation. Even though medical systems have been computerized in the past decades, the unprecedented nature of COVID-19 has brought forward issues that need to be automated but are still dependent on human intervention. For example, patient care and scheduling medical personnel are mostly done through a manual rule-based methodology to ensure that the people that are more infectious are handled with rigorous spread prevention measures. 62 The application of AI for such novel issues poses certain challenges, but overcoming these challenges and exploiting the available methodologies could result in a marked improvement in a variety of fields, especially within the medical domain. Table 3 lists a few COVID-19 related issues for which AI 9 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 24, 2020. ; https://doi.org/10.1101/2020.11.22.20232959 doi: medRxiv preprint may be an ideal candidate. One major drawback of traditional statistical techniques is their inability to take advantage of Big Data. For example, statistical techniques have often been used to find the most influential factors that leads to COVID-19 related deaths, 36 and the socio-economic issues that arise as a result of COVID-19. 63 While statistical studies are helpful, identifying specific clusters of features will be instrumental for decision making. Deep clustering techniques based on convolution neural networks have outperformed traditional methods 64 for such clustering problems. In addition to handling complex input data, the convolutional models learn spatial information and correlation that results in better performance. Similarly, traditional techniques do not handle Big Data efficiently. For example, contact tracing through the use of mobile application such as ''WeChat'' has been effective in many countries. 65 However, one major limitation is their inability to ingest a large volume of data for decision making and adapt accordingly. 66 Due to the improvement in processing power of mobile devices, AI algorithms can be adapted to solving problems such as proximity identification and improve the accuracy and reliability of these systems through adaptation and real-time learning. Furthermore, classical techniques often fail to model the spatio-temporal relationships of the input data. Studies have been conducted to find the correlation between the traveling patterns of individuals and how it correlates to the spread of disease. 65 However, individual-level tracking only focuses on geographical location and overlooks other factors such as social clusters, race and age. It would be interesting to consider these social clustering parameters for this study, as they may have significant correlation with the spread of pandemic. Graph Neural Networks (GNN) can be used to model and identify specific events in the spatial and/or temporal domain. The spatial information can be modelled through graphs and the information can be propagated within the nodes through message passing. In addition to GNNs, using state-of-the-art NLP techniques like Transformers, 67 spatio-temporal information that pays specific attention to events (e.g. sudden spikes in patients) in the temporal and/or spatial domain can be efficiently modeled. Despite the large number of recent vision-based solutions for COVID-19, 68 the risk level assessment for a particular environment is neither defined nor addressed properly. Comparing different environments (using a combination of vision techniques for social distancing, face mask identification and other visible metrics) and quantifying the risk of transmission in those particular environments would be helpful in making decision such as implementing social distancing measures and easing/strengthening problem. The rapid development and growing research in AI is amassed as preprints rather than peer reviewed work. Thus, the exclusion of preprints limits the scope of the review process. In addition to the exclusion criteria, . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this this version posted November 24, 2020. ; https://doi.org/10.1101/2020.11.22.20232959 doi: medRxiv preprint the setting of AI research in the contemporary world imposes restrictions to review. A large portion of annual investment (US$ 1·7 billion in 2018 69 ) for R&D in the intersection of AI and healthcare comes from tech corporations such as Google and Facebook. 70 Since a major portion of the AI research and development ends up in commercial deployment, they are not made available to the public. It is important to note that our recommendations in Section 4 were built on the premise that the methods we prescribed were not already explored by other researchers. Perhaps, the methods were implemented and not published because they did not yield the expected positive results. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 24, 2020. ; . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 24, 2020. ; https://doi.org/10.1101/2020.11.22.20232959 doi: medRxiv preprint No of papers about AI and COIVID . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 24, 2020. ; https://doi.org/10.1101/2020.11.22.20232959 doi: medRxiv preprint . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 24, 2020. ❖ Environmental [40] [41] [42] [43] ❖ Socio-economic [44] [45] [46] [47] [48] [49] [50] [51] ❖ Population density and income [57] ❖ Population density, health resources, and political [58] ❖ Geographical location, age, household structure, and medical [59] ❖ Population, migration index, GPD, and consumer metrics [60] ❖ Age, sex, race, economic status, and population flow [61] EXTRINSIC INTRINSIC DYNAMIC Figure 4 : Impact Modeling. The reference numbers of research papers are given in square brackets. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 24, 2020. ; https://doi.org/10.1101/2020.11.22.20232959 doi: medRxiv preprint . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 24, 2020. ; https://doi.org/10.1101/2020.11.22.20232959 doi: medRxiv preprint is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 24, 2020. ; is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 24, 2020. ; https://doi.org/10.1101/2020.11.22.20232959 doi: medRxiv preprint Find the correlation between COVID-19 and socio-economic problems. Classical methods do not consider complex feature sets. Deep clustering can be used as an alternative to identify complex features. Contact tracing through mobile applications. A major drawback is their inability to utilize Big Data efficiently. The increased processing power of mobile devices enables the use of AI to improve efficiency and performance. Forecast COVID-19 cases and deaths based on spatio-temporal data. Certain models fail to consider spatial and temporal relationship of data. Temporal information can be modelled effectively through Graph Neural Network and transformers. Find correlation between disease spread and population behavior such as travelling patterns. Classical techniques use individualistic data and do not consider the related features between individuals. Spatial information can be modelled using network models such as graph neural network to identify social clusters. Social distancing based on computer vision. Environment-specific risk assessment is an issue which has not been studied in detail. Existing CNN based solutions can be aggregated and extended for risk assessment. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 24, 2020. ; https://doi.org/10.1101/2020.11.22.20232959 doi: medRxiv preprint . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 24, 2020. ; https://doi.org/10.1101/2020.11.22.20232959 doi: medRxiv preprint . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 24, 2020. ; https://doi.org/10.1101/2020.11.22.20232959 doi: medRxiv preprint . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 24, 2020. ; https://doi.org/10.1101/2020.11.22.20232959 doi: medRxiv preprint . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 24, 2020. ; https://doi.org/10.1101/2020.11.22.20232959 doi: medRxiv preprint WHO Coronavirus Disease Artificial intelligence for global health Artificial intelligence in health care: enabling informed care Artificial Intelligence (AI) and Big Data for Coronavirus (COVID-19) Pandemic: A Survey on the State-of-the-Arts Imagenet classification with deep convolutional neural networks Artificial intelligence and the future of global health High-performance medicine: the convergence of human and artificial intelligence Artificial Intelligence against COVID-19: An Early Review Utility of Artificial Intelligence Amidst the COVID 19 Pandemic: A Applications of machine learning and artificial intelligence for Covid-19 (SARS-CoV-2) pandemic: A review Memory-based meso-scale modeling of Covid-19: County-resolved timelines in Germany Spatio-temporal resource mapping for intensive care units at regional level for COVID-19 emergency in Italy Risk of a second wave of Covid-19 infections: using artificial intelligence to investigate stringency of physical distancing policies in North America Time series forecasting of covid-19 transmission in canada using lstm networks Prediction and analysis of COVID-19 positive cases using deep learning models: A descriptive case study of India Time series forecasting of covid-19 using deep learning models: India-usa comparative case study Predictions for covid-19 with deep learning models of lstm, gru and bi-lstm Deep learning methods for forecasting covid-19 time-series data: A comparative study Prediction for the spread of COVID-19 in India and effectiveness of preventive measures Spatio-temporal estimation of the daily cases of covid-19 in worldwide using random forest machine learning algorithm Epidemiology of Coronavirus COVID-19: Forecasting the Future Incidence in Different Countries Forecasting the prevalence of COVID-19 outbreak in Egypt using nonlinear autoregressive artificial neural networks Forecast and evaluation of covid-19 spreading in usa with reduced-space gaussian process regression SutteARIMA: Short-term forecasting method, a case: Covid-19 and stock market in Spain Real-time forecasts and risk assessment of novel coronavirus (COVID-19) cases: A data-driven analysis Modified SEIR and AI prediction of the epidemics trend of COVID-19 in China under public health interventions Forecasting COVID-19-Associated Hospitalizations under Different Levels of Social Distancing in Lombardy and Emilia-Romagna, Northern Italy: Results from an Extended SEIR Compartmental Model Spatial epidemic dynamics of the COVID-19 outbreak in China Spatio-temporal patterns of the 2019-ncov epidemic at the county level in hubei province, china Spatiotemporal transmission dynamics of the COVID-19 pandemic and its impact on critical healthcare capacity. Health & place The COVID-19 pandemic: diverse contexts; different epidemics-how and why? Limiting the spread of COVID-19 in Africa: one size mitigation strategies do not fit all countries. The Lancet Global Health Global, regional, and national estimates of the population at increased risk of severe COVID-19 due to underlying health conditions in 2020: a modelling study. The Lancet Global Health Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study COVID-19 in Children in the United States: Intensive Care Admissions, Estimated Total Infected, and Projected Numbers of Severe Pediatric Cases in 2020 Estimating the infection and case fatality ratio for coronavirus disease (COVID-19) using age-adjusted data from the outbreak on the Diamond Princess cruise ship Estimates of the severity of coronavirus disease 2019: a model-based analysis. The Lancet Infectious Diseases A Spatio-Temporal Analysis of the Environmental Correlates of COVID-19 Incidence in Spain Prioritizing and Analyzing the Role of Climate and Urban Parameters in the Confirmed Cases of COVID-19 Based on Artificial Intelligence Applications A spatio-temporal analysis for exploring the effect of temperature on COVID-19 early evolution in Spain Neural network based country wise risk prediction of covid-19 Spatial analysis of COVID-19 clusters and contextual factors in New York City Relationship between Socio-Demographics and COVID-19: A Case Study in Three Texas Regions Assessing differential impacts of COVID-19 on black communities Spatiotemporal pattern of COVID-19 and government response in South Korea (as of May 31, 2020) Assessing the effect of containment measures on the spatio-temporal dynamic of COVID-19 in Italy and The Global Dynamic Interventions Strategies for COVID-19 Collaborative Group. Dynamic interventions to control covid-19 pandemic: a multivariate prediction modelling study comparing 16 worldwide countries Risk of a second wave of covid-19 infections: using artificial intelligence to investigate stringency of physical distancing policies in north america Evaluating the effect of city lock-down on controlling covid-19 propagation through deep learning and network science models Distribution of the COVID-19 epidemic and correlation with population emigration from Wuhan The COVID-19, Migration and Livelihood in India A Background Paper for Policy Makers International Institute for Population Sciences Population flow drives spatio-temporal distribution of covid-19 in china Examining the association between socio-demographic composition and COVID-19 fatalities in the European region using spatial regression approach Real-time forecasts of the COVID-19 epidemic in China from Modeling the Impact of Demography on COVID-19 Dynamics in Hubei and Lombardy Spatial statistics and influencing factors of the COVID-19 epidemic at both prefecture and county levels in Hubei Province Deepcovidnet: An interpretable deep learning model for predictive surveillance of covid-19 using heterogeneous features and their interactions Containing COVID-19 in the Emergency Department: The Role of Improved Case Detection and Segregation of Suspect Cases Covid-19 pandemic and mental health consequences: systematic review of the current evidence Deep clustering with convolutional autoencoders A New System for Surveillance and Digital Contact Tracing for COVID-19: Spatiotemporal Reporting Over Network and GPS A Survey of COVID-19 Contact Tracing Apps Attention is all you need A comprehensive survey of enabling and emerging technologies for social distancing-part ii: Emerging technologies and open issues Private Equity Investment in Artificial Intelligence New analysis shows important increases in investments in AI start-ups Real-time forecasts and risk assessment of novel coronavirus (COVID-19) cases: A data-driven analysis The authors have no conflicts of interest to declare.