key: cord-0062555-ifjdfm1j
authors: Yadav, Saneh Lata; Dhaiya, Ritika; Bhatia, Surbhi
title: Conclusions
date: 2021-04-30
journal: Researches and Applications of Artificial Intelligence to Mitigate Pandemics
DOI: 10.1016/b978-0-323-90959-4.00006-7
sha: 08a82671b6065ec813a1dfb968aecd8ce6e5b09f
doc_id: 62555
cord_uid: ifjdfm1j

This chapter presents the usage of data science, which further helps in exploring the global pandemic COVID-19. This disease suppresses an overwhelming burden, not only to healthcare systems but to the world's economy too. In this era of techniques and technologies, it is believed that data science can better utilize scarce healthcare resources. In this chapter, we provide an introduction of data science and its applications, which helps in combating different aspects of COVID-19. Publicly available datasets related to disease are used as community resources. Different kinds of datasets are used to analyze various aspects of pandemic at different scales. These different kinds of datasets can be audio, video, textual, speech, and sensor data. More than hundreds of research articles are also studied to prepare a bibliometric study. Apart from grabbing all the advantages from datasets, this paper highlights a few challenges, such as surety of correct data, need of multidisciplinary collaboration, new data modality, security issues, and availability of data.

discussed in Section 6.3. Bibliometric analysis for research in COVID-19 is elaborated in Section 6.4. Section 6.5 presents cross-cutting challenges. The chapter is concluded with a short summary.

Machine learning, statistical learning, time series modeling, expert system, data visualization, and probabilistic reasoning are few terms, which cover some applications of data science [4] . These all applications help in fighting with pandemics like COVID-19 in different manners. It helps in managing limited resources, developing plans, understanding the uncertainty, extracting information for building collaboration, etc. Few applications of data science in identifying disease are explained further.

Various risk assessment algorithms are opted to analyze different diseases like cancer, cardiac arrest, and diabetes. Artificial neural network (ANN) introduced a number of risk assessment algorithms [5] [6] [7] . Such algorithms take patient data like age, gender, traits, symptoms, state, and stage, for efficiently analyzing or estimating the risk. But this is a bitter truth that whenever a pandemic arises, medical resources fall into scarcity. Medical resources are lean globally for any pandemic situation. So, it is the need of time and situation to efficiently and effectively utilize the limited resources. Resources in peak times are utilized rapidly to manage patient's risk. Patient prioritization data can be prepared to sustain risk. Priority constraints can be different or as per situation to control the risk level.

COVID-19 spreads due to lack in screening and diagnosis. From China to worldwide, it majorly spreads due to ineffectual screening facilities at airports. In the initial days of pandemic, social distancing was a challenge for everyone and hence was not taken on serious mode. Mild symptoms were often ignored that became major reasons for the enormous spreading rate. There exist various remote diagnosis tools available, which help in self-analyzing the symptoms and assist in controlling the spread rate. To selfanalyze, various mobile applications are available, which use audio, video, and sound data to perform diagnosis. Such tools are helpful in diagnosing the symptoms when healthcare resources are lean. Automated tools are also encouraged at airports or at places where movement of people is nonstop [4] . For example, infra-red sensor-based temperature scanner.

Any modeling is required to manage capacity in an effective manner. So, as per a survey, to manage epidemic, compartment models are more popular [8] . In such modeling, people are divided into compartments using a simple differential equation. The SEIR model is also used to model COVID-19 spread, where four states are assumed to model the flow of people like susceptible, exposed, infected, and recovered [9] . In generative modeling, few circumstances are observed like effects of social distancing and causes of herd community. All these models must be updated from time to time to capture the best dataset [10] [11] [12] . Some use-case data are also developed to evaluate more accurately "what if '" kinds of situations. Table 6 .1 is organized to showcase a few epidemic models and its working features.

In few countries, it is opted in the early stage of COVID-19, to trace the contact person of infected one and send them to quarantine [11, 12] . This approach is very effective and helpful in controlling the spread rate. Various mobile phone applications are proposed to quickly identify and trace the contacted person. Automated diagnosis, online surveys, and smartphone contact sensing are few measures. These measures are useful to create alert to hospitals and government, with regard to the outbreak [16] . It is also supported in logistic planning like masks, gowns, sanitizers, test kits, hospital beds, and ventilators. For such planning and management, machine learning and data science algorithms do a good job. It predicts the need of ventilators and beds in hospitals in advance so that preparation can be done on an early basis. In peak times of pandemic, shortage of not only medical resources but also healthcare workers is triggered. To mitigate this, data science introduces Social media model Analyses on the bases of ambulance call out data and similar data [13] [14] [15] automated patient care tools. These tools can provide information about the outbreak, symptoms, and personal precautionary measures. WHO introduced an interactive chatbot to monitor health conditions in emergency cases [17] [18] [19] [20] , [21, 22] . Data science proves its eligibility in discovering drugs and treatment through Bayesian clinical trial methods, which works on collected data. It helps in identifying eligible patients to apply to clinical trials, hence reduce time spent in examining data, predicting protein structure, and genomes. Therefore, we can say that the above fields are popular where data scientists can contribute a lot [23] [24] [25] [26] [27] [28] [29] [30] [31] .

Governments have taken various steps to control the spread rate of CO-VID-19. A few use cases are highlighted and briefed further. 1. Monitoring social distancing: To monitor social distancing, governments adopt various strategies [32] [33] [34] . However, this is a nonpharmaceutical method to control COVID-19 but very effective. Natural language processing (NLP) is used to monitor social media information for analyzing social distancing. In many areas, air pollution level is used to analyze the rate of movement of people [35] . Cellular data are also traced to monitor the mobility rate. It is accepted by all that social distancing affects the global economy as well [36, 37] . It is a challenge for few organizations to get back to their previous level. For such cases, data science can be helpful in identifying optimal economic interventions. Organizations can detect unusual patterns of behaviors in the market [38, 39] .

Spreading false information is always harmful. Internet websites manage to maintain an up-to-date list of false information regarding COVID-19. Such false information can affect datasets also. The consumption of garlic and alcohol can get rid of COVID-19 is one such false information. Such information can harm public health. So, data science could be used to classify such misinformation. Machine learning techniques support sentiment analysis, which can help the public to be aware about false information [40] [41] [42] [43] [44] , [22, 45] .

Almost all the countries are tracing per day growth rate on different geographical locations [46] . The data are collected under different columns like daily positive cases, total positive cases, cured cases, and mortality cases. These data are further attributed as patient location, reporting data, past history of patient, and symptoms. India is also compiling data statewise then district wise, which is very useful in maintaining up to date datasets. Data visualization and predictive analysis, can be done on collective datasets for efficient measures. But data science would not work properly, if there is divergence in testing regimes. In few countries, test kits are inadequate, due to which accurate data are not available. So, it is a challenge for data scientists to work in such an environment [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] .

Nowadays, people are discussing and sharing textual data on social media platforms [67] . The major keywords like COVID-19, N95 masks, pandemic, and coronavirus aid in monitoring the spread rate of COVID-19 [54] . NLP also works here to monitor people's reactions on lockdown, social distancing, and personal hygiene. NLP measures the sentiments of people on social media and the Internet [68] . The role of data science is to record social media streaming that contains the above-mentioned keywords. Apart from social media data, academic publications are also flooded these days to provide textual data [55] [56] [57] [58] [59] [60] [61] . Such bibliometric datasets are available online. Many open research article datasets are already released by publication houses to motivate research in this field. WHO is working to provide up-to-date scientific literature for COVID-19. Table 6 .2 collects data for organizations, which are compiling scientific dataset on COVID-19 (freely available) [69] . 

Biomedical data sources are physical medical reports (X-ray reports of an individual) and clinical pathology reports (Genomic structure) [70] [71] [72] . These are helpful in diagnosis, prognosis, and treatment. Diagnosis means to check about a particular illness, prognosis means to predict the outcome of a disease and chances of recovery. So, for diagnosis and prognosis, human interpretation is required [73, 74] . A number of mobile apps are developed to self-diagnose and self-prognosis. Various automated diagnosis apps are developed to test chest X-ray, lungs computation tomography (CT) scan. The X-ray scans the chest and provides results on the basis of analysis from stored datasets. Some X-ray datasets are publicly available, which contain patient's details like date, demographics, findings location, survival information, and treatment. This data are analyzed on some neural network trained model [75] [76] [77] [78] [79] . However, these models are not self-sufficient to identify the problem, as it requires clinical experts to label reports. Another biomedical data source is clinical pathology reports, which contains datasets of genomic sequencing. The genomic dataset includes drug impact, protein-protein interaction, and RNA structures, which support diagnosis test evaluation. Another challenge is availability of datasets to accurately measure problem statements. Same issue arises with lung CT scan, as public availability of biomedical dataset is very difficult [80] [81] [82] [83] .

Some other relevant datasets are air quality index statistics, and datasets for mobility of people, which monitors factors related to COVID-19. If the mobility of people is controlled, it automatically affects air quality or pollution level. After a study, it has been noticed that few popular cities' air quality has improved in recent days [35, 84, 85] . Google has also released datasets after mobility tracing, which is compiled through Google maps.

Kaggle has organized various competitions to promote and facilitate research in data science. The participants use publicly available datasets and recent research articles on COVID-19. Such competitions are motivated by announcing cash prizes also.

The previously discussed datasets were publicly available and have some limitations. So, in the next section, some ongoing research ideas are discussed.

Computer vision technology plays a vital role in detecting disease by tracing image dataset. This technology sped up the process of disease detection and proved its potential by outperforming expert radiologists. For COVID-19, two computer vision modalities are used. One is CT scan and another is X-ray scan. Radiologists diagnose many COVID-19 patients through chest CT scan. It is a very successful attempt because nowadays machine learning and deep learning techniques are integrated with CT scan technology, which are useful in detecting radiographical changes in patients very frequently [86] [87] [88] . Initially, it was evaluated and tested upon many COVID-19 positive cases to improve its reliability and trust factor [89-100].

Pneumonia and cough are the common symptoms of COVID-19. Change in voice due to pneumonia and cough can be taken as audio/sound analysis [101] [102] [103] [104] [105] [106] [107] . This analysis can be done with low-cost smartphones too. The speech pattern and cough pattern are the main data source to monitor the situation [108] . AI4COVID-19 is a popular app developed by preliminary diagnosis. It collects data for sound for future dataset observations and processes the feasibility of detection with 90% promising rate.

Sensors are the tiny embedded devices, which sense the environment before processing. Sensors can be deployed anywhere in remote areas as well. In the medical field, sensors can be deployed in a patient's body to diagnose the variations of body glucose, temperature, blood pressure, heart rate, pulse rate, etc. [109] . Sensors can provide demographic data, mobility data, and disease related data and user generated data from social media. Such a system is named as α-Satellite, which can assess risk level [110] . The data fetched from this system can be used to diagnose COVID-19. This system proves its reliability because it uses multireading to sense the symptoms of disease. Smartphones embedded with sensors can detect movement of people during pandemic. The table below has sensor based systems developed for pandemic with their specific features (Table 6 .3).

The above-mentioned apps are developed by different countries to provide safe and sound surroundings to their people. These apps were developed for the safety of people but users face few privacy issues [116] [117] [118] [119] . Many of the apps require uploading the contact list, which can reveal nationwide databases. To minimize such weakness, Pan-European Privacy-Preserving Proximity Tracing Consortium (PEPP-PT) developed an app named Decentralized Preserving Proximity tracing (DP-3T) [120, 121] . It provides a preservation alert to the users who may have been in contact with an infected person. Another similar app launched with homomorphic encryption features. Google and Apple have also announced to develop a privacy preserving contact tracing app based on Bluetooth.

Mobile Technologies: Mobile technologies are being used for a variety of purposes in healthcare. Most importantly, they are enabling new ways for pandemic management by providing powerful tools to both doctors and patients for effective prevention and treatment [84, 85, 122] . As the common risk factor of pandemic are related to human behavior, therefore, mobile phone-based health solutions can be used to combat the rising burden of pandemic by focusing on behavioral change programs to promote a healthy lifestyle. This chapter discusses the common pandemic, their burden, and future estimated projections and shows how mobile phone technologies can provide effective pandemic management in developing countries, which have a lot of issues in their healthcare systems.

Some researchers put extensive efforts in discovering new drugs to support SARS-COV2 [123] . To build a model to explore 3D structure of SARS-COV2, AlphaFold model has been developed, which is a deep learning based model. This model is based on dilated ResNet architecture that predicts the distance and the distribution of angles between acid residing on protein structure [124, 125] (Table 6 .4). StayHomeSafe [111] Track for your safety 5.

Home Quarantine app [112] Detect if quarantines rules have been obeyed or not 6.

Close Contact Detector [113] Track the infected contact persons 7.

Track Together [114] Track the infected contact persons 8.

HaMagen [115] Track the infected contact persons

After developing drugs, it should undergo clinical trials before deployment to prove its effectiveness. For such trials, randomized clinical trials (RCTs) were good. But RCT fails to prove its effectiveness for elderly patients and patients at higher risk. So, few improvements were done with RCT to demonstrate its efficiency and effectiveness. These improvements are done by integrating ML applications. Previously, in RCT trials, patients for treatment are allocated in uniform randomization, which can be highly suboptimal in terms of learning. In improved RCT, patients are first observed before sending clinical trials. It speeds up the process and has significantly reduced error rate. After successful clinical trials, data finally received are certain for targeting particular treatment. For more improvements in drug discovery, WHO, European Medicine agency, UK medicines, healthcare products regulatory, and US food are drug administration has established accelerated clearance pathways. 534 clinical trials were taken till March 24, 2020. But, a challenge in clinical trials is recruiting suitable patients [127] . So, data-driven solutions are best to identify eligible patients who have gone through remotely monitored checkups (Table 6 .5). Predict potential lead compounds targeting SARS-COV2 6.

Collaborative and Antiviral Discovery Model

To discover molecules to fight against COVID-19 

Researchers are working on writing articles for COVID-19 to spread awareness. It helps in improving data repositories. Articles are peer-reviewed and nonpeer-reviewed. Peer-reviewed articles are crawled by scopus and nopeer reviewed articles are crawled by arXiv, medRxiv, etc. [128] [129] [130] [131] [132] [133] . The most popular sources for peer-reviewed article includes The Lancet with more than 228 articles, Nature with more than 204 articles and many more are arranged in table [134] [135] [136] [137] . The dataset includes title of article, author details, journal name, publication date, etc. Basically, this dataset is extracted by keyword matching technique. The popular keywords used for extraction are CoronaVirus, COVID-19, COVID, Epidemic, Pandemic, SARS-COV2. Manual verification has also been done after keyword matching extraction to avoid not related data. The dataset covers more than 3500 publications till our study June, 2020. Most peer-reviewed and nonpeer-reviewed articles are written by Chinese researchers. After China, United States researchers contributed a lot. The pandemic situation has resulted in rapid production of academic data. Peer-reviewed journals are less in number as compared to nonpeer-reviewed journals, due to the urgency of dissemination. For more information and data, researchers are looking over preprint articles. In the end it is concluded that the rate of publication for COVID-19 is growing faster if compared with some other past epidemic like Ebola, SARS-COV, and MERMS. More than 1000 peer-reviewed publications have been recorded in around 3 months. SARS-COV1 and Ebola have reached this count in 3 years. COVID-19 is getting more attention in academics for research (Tables 6.6 and 6.7).

For machine learning models to work, they need to be fed with high fidelity and voluminous data. Data become a limitation in that sense. For use cases such as "speech analysis, extensive labeled datasets are not around yet. However, for a few other use cases such as textual analysis and medical images, the datasets needed are smaller as compared to the ones needed for deep learning models [69, [138] [139] [140] [141] [142] . The distributed data sources being distributed contributes to measured data being scarce. For example, if we talk about electronic healthcare records, we know that these are often bucketed into several sections at national, regional, or hospital level, which brings us to challenge in getting consistent and measurable data, in terms of schemas. To overcome these barriers, automation algorithms for data wrangling, munging processes will become critical [122, 143] . Apart from these challenges around data being available, we often see challenges with the available data as well. Due to the research being extremely time-critical, it is been hard to create reliable datasets. For example, social media data can quickly become out-to-date from the time it is captured to the time it is put into a usable format. As a result, real-time datasets are riddled with poorly quantified biases. Analytical approaches to tackle these data challenges can be an area of exploration, however, not an easy solution.

It is a very high time to indulge in research for COVID-19. However, most of the methods that are suggested in this chapter are based on datasets and derived statistical outcomes. It is important to consider that the research outcomes can affect healthcare policy. For instance, these could be supported by several governments to come up with social distancing policies. Although people in these decision maker positions may not have the To capture uncertainty of results, Bayesian methods can be used, though we have not come across many quantified studies until now [144] . Reproducible conclusions are further necessary to make sure that data analysis was conducted correctly. This task will further compound the challenges. "Explainable AI" is another route that can be explored to tackle this. However, a caveat must be added here that there is not complete confidence in whether this will guardrail against issues such as unintentional bias or adversarial scenarios.

As we start exploring the research suggestions, protecting privacy and adhering to ethical standards will become paramount. This will directly impact how much scale we can reach in adoption across populations, as infrastructure setup may continue after the pandemic. Efforts are being made around building medical analytics that can preserve privacy.

Floridi et al., outline some consensus around five major artificial intelligence ethics principles, namely (1) beneficence, (2) nonmaleficence, (3) autonomy, (4) justice, and (5) explicability. However, COVID-19's unique situation may make it tricky to balance these AI ethics virtues. Other questions that are unanswered at this point relate to the allocation of scarce resources and the tradeoffs in it. Call of action presented by a group of experts on data governance highlights the need to share data between public and private sectors to ensure that data are used for "beneficence" where it is needed and prevent maleficence [145, 146] .

Furthermore, privacy will become important as we start rolling out the interventions, which may have sensitive data (e.g., targeted social distancing measures). In this regard, simple steps can be taken to ensure ethical data science research. Data collection should be transparent (informing the users about the data being collected).

COVID-19's long-term impact is still unknown. A mix of domain expertise from multiple fields is needed to draw insights, along with international collaboration and tracking of COVID-19. For instance, black-box models may result in a practical solution, but that solution could be useless with-out the involvement of (international) medical and biotechnology expert interpretations. Hence, bringing together many cohorts of complementary expertise becomes useful, which presents new challenges, such as ethics, benefits, and risks, that are clearly articulated.

Certain data modalities that can have a big role to play are not readily available for research. A huge challenge comes around adaptation of existing techniques to reflect new data types. For instance, the data science community has established good expertise in computer vision tasks, however, if we talk about processing ultrasound scans, there is a lot that is yet to be figured out. They also bring benefits around greater ease of use, absence of radiation while being a low-cost solution. Despite these known advantages, to the best of our knowledge, no study has yet explored the potential of automatically detecting COVID-19 infections via ultrasound scans. Similarly, magnetic resonance imaging (MRI) is considered the safest imaging modality as it is a noninvasive and nonionizing technique, which provides a high resolution image and excellent soft tissue contrast. Some studies such as touch upon the significance of MRI in fighting against COVID-19 infections. Yet, lack of sufficient training data restricted exploration of the data modalities. Thus, a challenge is to rapidly develop a well-annotated dataset of such medical imaging modalities.

In lots of developing nations, large populations have limited access to healthcare and information, which created unique challenges related to COVID-19 pandemic. Technology can help solve these challenges; however, the scale can be quite challenging in making it globally inclusive. The models should consider the applications in rural as well as economically deprived regions. For example, when creating a contact tracing app, a few things to be kept in mind are: is it low-cost, how many resources does it need to be used, can it be used with limited network connectivity, does it support multiple languages, is it accessible to illiterate users or those with disabilities, etc. To address this global pandemic, we must ensure emphasis on widespread accessibility of technological solutions.

This chapter has been written to rapidly make available a summary of ongoing work for the wider community. It talks about how artificial intelligence has been used to tackle many aspects of the COVID-19 crisis at different scales including molecular, clinical, and societal applications. AI including machine learning has found many applications in understanding challenges created by COVID-19 in the medical and societal realm. Most of them, however, are in the nascent stages and it will take some time before we can show how these can create impact at scale. Let us cover each of these challenges. At the molecular scale, biochemistry applications of AI can be used to understand the proteins' structure of SARSCoV-2. Along with that, AI can be used to discover how existing drugs may be effective against the virus as well as find new compounds to make potential drugs or potential vaccines. It can also be used to improve our understanding about the virus and help improve diagnosis. At the clinical scale, AI can be applied in medical imaging to screen and diagnose the virus, while exploring alternate ways to find the disease via noninvasive devices such as mobiles. It can also help to predict prognosis of the patients by utilizing various data inputs. At the societal scale, AI applications around epidemiological research modeling empirical data have been used to forecast COVID-19 stats data such as number of cases, mortality and recovery rates. Along with that, AI has been used to look for patterns of similarities and differences in the evolution of the pandemic across regions. AI can also be used to analyze COVID-related content across social media to make sure the right kind of information is shared, and incorrect or misleading information can be controlled. To continue efforts in these directions, it becomes extremely imperative that scalable sharing and hosting of datasets and models is made possible. This will help understand where AI can be of value against the pandemic. AI targeting biomedical applications around clinical and molecular data should include direction from regulatory and quality frameworks to minimize potential risks while ensuring the validity of usage. Along with the data considerations, at the global level, international AI cooperation will become paramount as we begin research adoption to create solutions at the global scale, which can then be applied to local contexts and situations.

Clinical features of patients infected with 2019 novel coronavirus in

A novel coronavirus from patients with pneumonia in China

Mapping the landscape of artificial intelligence applications against COVID-19

Leveraging Data Science To Combat COVID-19: A Comprehensive Review

Artificial intelligence for cancer-associated thrombosis risk assessment-Author's reply

Artificial intelligence based clinical risk assessment in predicting cardiac related chest pain in patients presenting to emergency room

Mobile technologies for managing non-communicable diseases in developing countries, Mobile Applications and Solutions for Social Inclusion

Early dynamics of transmission and control of COVID-19: a mathematical modelling study

Dynamic causal modelling of COVID-19

About coronavirus disease 2019 (COVID-19)

Impact of non-pharmaceutical interventions (NPIs) to reduce COVID19 mortality and healthcare demand

Modeling infectious disease dynamics in the complex landscape of global health

Characterizing the propagation of situational information in social media during covid-19 epidemic: a case study on weibo

Informational flow on Twitter-Corona virus outbreak-topic modelling approach

Predicting commercially available antiviral drugs that may act on the novel coronavirus (SARS-CoV-2) through a drug-target interaction deep learning model

Feasibility of controlling COVID-19 outbreaks by isolation of cases and contacts

Quantifying SARS-CoV-2 transmission suggests epidemic control with digital contact tracing

Identification of COVID-19 can be quicker through artificial intelligence framework using a mobile phone-based survey when cities and towns are under quarantine

Coronasurveys: Monitoring COVID-19 incidence via open polls

EpiMap: towards quantifying contact networks for understanding epidemiology in developing countries

The covid-19 social media infodemic

A first look at COVID-19 information and misinformation sharing on Twitter

WHO and Rakuten Viber fight COVID-19 misinformation with interactive chatbot

Health Care Workers' Stress Compounded By Long Days And Concerns About People Not Taking COVID-19 Seriously

Artificial intelligence in pharmaceutical research and development

Artificial intelligence in drug development: present status and future prospects

Artificial intelligence for drug discovery, biomarker development, and generation of novel chemistry

An agent-based approach to real-time patient identification for clinical trials

Bayesian clinical trials

Isolation, quarantine, social distancing and community containment: pivotal role for old-style public health measures in the novel coronavirus (2019-nCoV) outbreak

A simplified model for expected development of the SARS-CoV-2 (Corona)

Can Twitter predict disease outbreaks?

World's air pollution: Real-time air quality index

The global macroeconomic impacts of COVID-19: Seven scenarios

What will be the economic impact of covid-19 in the us? rough estimates of disease scenarios (No. w26867)

The use of Twitter to track levels of disease activity and public concern in the US during the influenza A H1N1 pandemic

Early evidence that COVID-19 government policies reduce urban air pollution

How to fight an infodemic

See something, say something: correction of global health misinformation on social media

The spread of medical fake news in social media-the pilot quantitative study

Misinformation related to the 2019-20 coronavirus pandemic

Fighting COVID-19 misinformation on social media: Experimental evidence for a scalable accuracy-nudge intervention

Covid-19 and computer audition: an overview on what speech & sound analysis could contribute in the SARS-CoV-2 Corona crisis

China's QR health code

CSSEGISandData/COVID-19

Novel corona-virus dataset

European Centre for Disease Prevention and Control (ECDC)

COVID-19 Hospital Impact Model for Epidemics

COVID-19 Korea Dataset with Patient Routes

Genomic epidemiology of hCoV-19

Covid-19: The first public coronavirus twitter dataset

Coronavirus (covid19) tweets

CORD-19 research challenge

Keep up with the latest coronavirus research

Global research on novel coronavirus2019

COVID-19 Community Mobility Reports

COVID-19 DATABASE

Dataset of infections in Germany

Covid-19-BSTI Imaging Database

ncov2019 epidemiological data

COVID-19 image data collection

RCSB Protein Data Bank: Enabling biomedical research and drug discovery

Understanding the covid19 outbreak: a comparative data analytics and study

Improving the performance of cough sound discriminator in reverberant environments using microphone array

New york times dataset

Twitter catches the flu: detecting influenza epidemics using Twitter

Computer Vision for X-Ray Testing

CT scan dataset about CO-VID-19

Coronada

A fully automatic deep learning system for COVID-19 diagnostic and prognostic analysis

Very deep convolutional networks for large-scale image recognition

bilenets: Efficient convolutional neural networks for mobile vision applications

Automatic detection of coronavirus disease (covid-19) using x-ray images and deep convolutional neural networks

Deep residual learning for image recognition

Covid-resnet: a deep learning framework for screening of co-vid19 from radiographs

Chemml: A machine learning and informatics program package for the analysis, mining, and modeling of chemical and materials data

Digital image analysis in breast pathology-from image processing techniques to artificial intelligence

Using artificial intelligence to read chest radiographs for tuberculosis detection: A multi-site evaluation of the diagnostic accuracy of three deep learning systems

Deep learning in chest radiography: detection of findings and presence of change

Mobile phone data and COVID-19: missing an opportunity?

Data world

Correlation of chest CT and RT-PCR testing in coronavirus disease 2019 (COVID-19) in China: a report of 1014 cases

COVID-19): role of chest CT in diagnosis and management

A deep learning algorithm using CT images to screen for Corona Virus Disease (COVID-19)

Covidx-net: A framework of deep learning classifiers to diagnose covid-19 in x-ray images

Covid-19 screening on chest x-ray images using deep learning based anomaly detection

Coronavirus (covid-19) classification using ct images by machine learning methods

Coronavirus detection and analysis on chest CT with deep learning

Automated methods for detection and classification pneumonia based on x-ray images using deep learning

Classification of COVID-19 in chest X-ray images using DeTraC deep convolutional neural network

Can AI help in screening viral and COVID-19 pneumonia?

COVID-2019 Detection Using X-Ray Images And Artificial Intelligence Hybrid Systems

Covid-19 detection using artificial intelligence

Covid-xpert: an AI powered population screening of covid-19 cases using chest radiography images

Covid-19: automatic detection from x-ray images utilizing transfer learning with convolutional neural networks

Using Twitter and web news mining to predict COVID-19 outbreak

Deep learning system to screen coronavirus disease 2019 pneumonia

Deep learning-based model for detecting 2019 novel coronavirus pneumonia on high-resolution computed tomography: a prospective study

Rapid ai development cycle for the coronavirus (covid-19) pandemic: Initial results for automated detection & patient monitoring using deep learning ct image analysis

Artificial intelligence distinguishes COVID-19 from community acquired pneumonia on chest CT

Diagnosis of pneumonia from sounds collected using low cost cell phones

Automated screening for distress: a perspective for the future

CAST a database: Rapid targeted large-scale big data acquisition via smallworld modelling of social media platforms

satellite: an ai-driven system and benchmark datasets for hierarchical community-level risk assessment to help combat COVID-19

Poland: App helps police monitor home quarantine, accessed on: 1-April-2020

Google uses location data to show which places are complying with stay-at-home orders -and which aren't, The Verge

A novel ai-enabled framework to diagnose coronavirus covid 19 using smartphone embedded sensors: Design study

Geographical tracking and mapping of coronavirus disease COVID-19/severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) epidemic and associated events around the world: how 21st century GIS technologies are supporting the global fight against outbreaks and epidemics

The economics of artificial intelligence

Israelis using voluntary coronavirus monitoring app

TraceTogether -behind the scenes look at its development process, published on 25

Coronavirus: UK considers virus-tracing app to ease lockdown

Contact tracing mobile apps for COVID-19: privacy considerations and related trade-offs

Health surveillance during covid-19 pandemic

D3tp documentation

Protein structure prediction using multiple deep neural networks in the 13th Critical Assessment of Protein Structure Prediction (CASP13)

Uncover covid19 challenge

What is the people posting about symptoms related to Coronavirus in Bogota Colombia?

Prediction of potential commercially inhibitors against SARS-CoV-2 by multi-task deep model

Global coalition to accelerate COVID-19 clinical research in resourcelimited settings

Potential COVID-2019 3c-like Protease Inhibitors Designed Using Generative Deep Learning Approaches

The use of Bayes and causal modelling in decision making, uncertainty and risk

Wikipedia database download

COVID-Net: a tailored deep convolutional neural network design for detection of COVID-19 cases from chest X-ray images

Towards detecting influenza epidemics by analyzing Twitter messages

Flu detector-tracking epidemics on Twitter, Joint European Conference on Machine Learning and Knowledge Discovery in Databases

Coronavirus on social media: analyzing misinformation in Twitter conversations

A comparative study on the clinical features of COVID-19 pneumonia to other pneumonias

Understanding human language: Can NLP and deep learning help?

The White House Office of Science and Technology

Contact tracing in the real world

Multi-agent system for recruiting patients for clinical trials

A novel approach for crawling the opinions from world wide web