key: cord-301035-dz8642qx
authors: Rasheed, Jawad; Jamil, Akhtar; Hameed, Alaa Ali; Aftab, Usman; Aftab, Javaria; Shah, Syed Attique; Draheim, Dirk
title: A Survey on Artificial Intelligence Approaches in Supporting Frontline Workers and Decision Makers for COVID-19 Pandemic
date: 2020-10-10
journal: Chaos Solitons Fractals
DOI: 10.1016/j.chaos.2020.110337
sha: 
doc_id: 301035
cord_uid: dz8642qx

While the world has experience with many different types of infectious diseases, the current crisis related to the spread of COVID-19 has challenged epidemiologists and public health experts alike, leading to a rapid search for, and development of, new and innovative solutions to combat its spread. The transmission of this virus has infected more than 18.92 million people as of August 6, 2020, with over half a million deaths across the globe; the World Health Organization (WHO) has declared this a global pandemic. A multidisciplinary approach needs to be followed for diagnosis, treatment and tracking, especially between medical and computer sciences, so, a common ground is available to facilitate the research work at a faster pace. With this in mind, this survey paper aimed to explore and understand how and which different technological tools and techniques have been used within the context of COVID-19. The primary contribution of this paper is in its collation of the current state-of-the-art technological approaches applied to the context of COVID-19, and doing this in a holistic way, covering multiple disciplines and different perspectives. The analysis is widened by investigating Artificial Intelligence (AI) approaches for the diagnosis, anticipate infection and mortality rate by tracing contacts and targeted drug designing. Moreover, the impact of different kinds of medical data used in diagnosis, prognosis and pandemic analysis is also provided. This review paper covers both medical and technological perspectives to facilitate the virologists, AI researchers and policymakers while in combating the COVID-19 outbreak.

First China, then worldwide [16] To tackle the pandemic of COVID-19, scientists and medical experts are trying to come up with state-of-the-art solution to control the spread, identify infected patients, monitor virus growth and develop vaccines. Many researchers are working with computer-based diagnosis techniques using clinical blood samples, radiography images, and respiratory-related datasets. AI and machine learning techniques are helping medical personnel in monitoring, analysis and prediction of various critical applications such as survival mortality assessment, forecasting models, virions sequence formation and drug discovery models.

In order to help stop the pandemic, the early detection of positive cases is crucial as it facilitates the containment and isolation of the virus, thus reducing its spread. Currently, the standard procedure followed for identification of COVID-19 positive cases is performed by Reverse-Transcription Polymerase Chain Reaction (RT-PCR) [22] . To further improve and hasten the early detection process, there is room for the development of new and innovative tools that may improve the early detection and tracking of the virus [23, 24] .

As there is a clear need for these innovative technical solutions, there has, logically, been a large amount of rapid-paced development of these tools. However, due to the speed and number of these developments, it is easy to become overwhelmed, thus limiting the ability to critically examine, explore, and understand what a given solution does, how it performs, and whether or not it may or may not be beneficial. Thus, with this in mind, the main objective of this comprehensive review is to provide an in-depth account of COVID-19 and various advancements in AI techniques that have been currently developed to help manage or combat COVID-19.

The review was conducted between July and August 2020, using the keywords -COVID-19‖ AND (combined with one of the following: -AI‖, -Machine Learning‖, -Deep Learning‖) as well as -COVID-19‖ AND -AI‖ AND (combined with one of the following: -Clinical Blood Samples‖, -Forecasting Models‖, -Drug Discovery Models‖). This approach is, admittedly, broad, but allowed for a large number of relevant papers to be identified. Additionally, due to the nature of this study being a relatively new research subject, pre-print papers (such as those submitted to arxiv) were not precluded. As recommended by [25] , we selected Google Scholar as our electronic bibliographic databases to remove any kind of biasness for any specific scientific publisher. We then excluded papers not written in English, duplicates, guest editorials, poster sessions, and blogs. Once the papers were identified, they were analyzed and grouped based on their aims and approaches.

Based on this conducted study, it is hoped that this paper will serve as a common ground for both medical experts and computer scientists to better understand the current state-of-the-art associated with COVID- 19 With this in mind, the main contributions of this study are as follows:

 Detailed history, characteristics, taxonomy, symptoms, behavior, and patterns of COVID-19 are outlined.

 Presentation and discussion of AI (both traditional ML and advanced DL) approaches applied to COVID-19 in a variety of aspects and domains.

 Description of major COVID-19 diagnosis techniques using clinical blood samples, radiography images, respiratory-related data for various critical applications such as survival mortality assessment, forecasting models, virions sequence formation and drug discovery models.

 A detailed discussion on issues, challenges, and recommendations to tackle the pandemic through AI and timely facilitate effective decision making is presented.

In order to meet these objectives, the rest of the paper is organized as follows. Section 2 lays out the historical overview, origin and means of transmission, and global dispersion of the virus. Section 3 highlights the novel AI techniques that have been applied to COVID-19, and furthermore, it explores non-applied, but potentially useful, AI-based applications for vaccine discovery, mortality and survival rate prediction, and outbreak forecasting. Section 4 discusses the challenges in COVID-19

research. Finally, the review is concluded in Section 5 by presenting a summarized overview of the conducted studies.

The type of coronaviruses that cause infections in humans belongs to a subfamily called Coronavirinae which is part of a larger [26].

The name Corona comes from the fact that the viral particle has spike-like projections or structures on its outer layer that can be seen under a scanning electron microscope. The RNA of coronaviruses is single-stranded, positive sense, has a diameter of around 80 to 120 nm and a nucleic acid length from 26 to 32 kbs [27] . They are classified under four variants or genera entitled as alpha (α), beta (β), gamma (γ) and delta (δ) [27] . It has been observed that mammals typically get infected by the α-and β-CoV while the other two (γ-and δ-CoV) mostly infect birds. From the α-CoV group, HCoV-NL63 and HCoV-229E have been found to affect humans, normally causing a mild respiratory illness, similar to the common cold. From the β-Cov group, SARS-CoV and MERS-CoV both tend to cause moderate to severe respiratory illnesses with a higher morbidity and mortality rate [28, 29] .

The most recent addition to the coronaviruses is that of SARS-CoV-2, which became known in late 2019 after a number of atypically presenting pneumonia cases were identified in Wuhan China; it was initially speculated that these cases might originate from the Huanan Seafood market [30] . The hallmark sign of many of the cases was the development of certain symptoms including dry cough with fever, breathing difficulty and a-typical presentation of pneumonia. On January 7 th 2020, Bangladesh have also registered a high number of cases; Figure 1 shows the top 15 countries for coronavirus infections and deathsbased on the latest WHO situation reports [34] . The disease itself is so pervasive due to the ease with which it is able to transmit itself, via small respiratory droplets (such as those created when talking, sneezing, coughing, and breathing) and has a higher level of danger to a person's health due to the current absence of targeted pharmaceutical solutions such as vaccines or antiviral drugs.

Pandemics are an ever-present threat to human society, health, and wellbeing and, therefore, attention must be dedicated to understanding how they occur, how they can be prevented, and how we can strategically mitigate the negative effects they cause. The current pandemic associated with COVID-19 is unique in that it represents the first and largest pandemic in modern times where new technological solutions, e.g. AI, are not only applicable to the pandemic, but able to be created by a large group of stakeholders from scientists to doctors to AI-hobbyists alike. As the pandemic has caused great disruption to normal day-to-day operations and created a sense of unknown amongst the public, many motivated scientists and citizens have tried to assist in the COVID-19 response by developing their own unique AI-based tools to solve a large number of problems, in a variety of applied domains, such as: COIVD-19 disease detection and classification, mortality rate prediction and severity assessment, outbreak forecasting and tracking, biological insight of SARS-Cov-2 strain, and drug discovery.

In order to explore the different approaches that are currently being developed and trialed, this section presents a selection of the different ways in which AI-based solutions have been applied to the COVID-19 pandemic. In order to provide a visual representation of how an AI-based COVID-19 system may work, a generic model has been created in Figure 2 . The system can process various types of raw data including images, blood samples, crowed sourced data, even GPS information for tracking positive cases, etc., which can be stored in a repository. Data mining techniques are usually applied to clean the raw data that is and then stored in a database for further processing and retrieval. ML and DL methods can be applied to this data for better visualization, diagnosis, and forecasting, which can help the end-users making better decisions and taking actions to mitigate the disease.

Early diagnosis of COVID-19 is important to ensure better outcomes for those who have been infected. At the beginning of the pandemic, due to the sense of the unknown that surrounded the new disease, one way that was being used to diagnose patients was via radiographic imagery (such as CT scans or X-Rays) of the lungs. Table II .

There has also been a large number of DL-based approaches proposed to identify COVID-19 from radiographic imagery in the literature. For example, Brunese et al. [47] exploited the transfer learning technique and fine-tuned DL-based modified Visual Geometry Group (VGG-16) model. The proposed framework was divided into 2 models that considered 6523 chest X-rays (2753 belongs to patients with pulmonary diseases, 250 of COVID-19 infected patients, while 3520 related to healthy patients).

The first model discriminates between healthy and pulmonary diseases patients, while the second model classifies the chest Xray (CXR) of pulmonary disease patient as COVID-19 or other pneumonia. In the end, if a CXR is classified as COVID-19, the framework provides a visualization of the CXR highlighting the potentially SARS-CoV-2 infected area. From the experimental data, it was observed that the discriminatory model (model 1) achieved accuracy, sensitivity, specificity and f-measure of 96%, 96%, 98% and 94% respectively, whereas the disease classification model (model 2) yielded an accuracy of 98%, sensitivity of 87%, specificity of 94%, and an f-measure of 89%.

Similarly, to identify SARS-CoV-2 infected cases from CXR images, various CNN frameworks such as VGG19, MobileNet v2, Inception, Xception, and Inception ResNet v2 have been evaluated using transfer learning by Apostolopoulos and Mpesiana [48] . From their conclusion, it was observed that MobileNet v2 outperformed the other tested state-of-the-art architectures by securing 99.10% sensitivity, 97.09% specificity, 97.40% accuracy on the two-class problem and 92.85% accuracy on the threeclass problem. In that paper, the models were trained and evaluated on a dataset of 1427 CXR images that constitutes 700 images of bacterial pneumonia infected patients, 224 images of COVID-19 infected cases, and 504 belonging to normal patients. MobileNet v2 also performed well yielding a sensitivity, specificity, two-class problem accuracy, and three-class Other identified DL-based methods and approaches are highlighted in Table III . 

Another type of COVID-19 diagnostic tool that has been developed is related to respiratory wave data, such as developed in Wang et al. [114] . In this paper, the authors proposed a respiratory pattern classification model that was based on a novel GRU neural network [115] , which extends GRU networks with bidirectional and attentional mechanisms (BI-AT-GRU) by using time-series real-world and stimulated respiratory data. The model detects and distinguishes the Tachypnea respiratory pattern, a more rapid respiration symptom that occurs in COVID-19 infected patients, from 6 other viral infection patterns. A novel Respiratory Stimulated Model (RSM) was used to generate stimulated breathing patterns to fill the scarce real-world data. The trained model was evaluated on real-world data captured by a depth camera. The proposed BI-AT-GRU model resulted in precision, recall, f1-score, and accuracy of 94.4%, 95.1%, 94.8% and 94.5% respectively. Table V lists the other identified COVID-19 AI-based applications that utilize respiratory or coughing data.

Outside of just diagnosis of COVID-19, AI-based tools have also been applied to identifying the severity of the disease in a patient, as well as to develop and optimize treatment strategies. For example, in order to attempt to help to reduce the mortality of COVID-19, and to optimize potential treatment strategies, a novel DL-based framework, combined with a multivariate logistic regression model, was proposed by Bai et al. [119] . Multi-layer perceptron (MLP) was used to convert each statistical sample containing 75 clinical data characteristics to a 40-dimensional feature vector. The proposed model was then evaluated using a dataset of 133 patients and achieved an overall accuracy of 89.1% and 95.4% AUC under 5-fold cross-validation repeated 5 times. The researchers found six key risk factors (age, lymphocyte, comorbid with hypertension, albumin, hypersensitive C-reactive protein level, and progressive consolidation in CT images) plus fibrosis in CT as a predictive factor for malignant progression.

Albahri et al. [120] also proposed an ML-based rescue framework that was combined with a novel multi-criteria decision- Chen et al. [121] developed a fatality rate predictive model for severe SARS-CoV-2 patients based on five distinct ML approaches, namely: elastic net, bagged flexible discriminant analysis (bagged-FDA), logistic regression, random forest and partial least squares regression. The model used data of 183 severely infected COVID-19 patients, out of which 115 survived and found four key features and clinical pointers (age, d-dimer level, lymphocyte count, and high-sensitivity C-reactive protein level) for survival/mortality assessment. Moreover, the developed models were evaluated on an external validation set that consisted of 64 severe COVID-19 confirmed cases, out of which 33 were survivals. The experimental results concluded that the logistic regression model was to be preferred over other ML models due to its high interpretability and simplicity; it obtained sensitivity, specificity and AUC of 89.2%, 68.7% and 89.5% respectively. Table VI shows other identified AI-based approaches related to severity and fatality assessment models.

One of the first areas where AI was applied to the COVID-19 pandemic was related to the development of outbreak forecasting models, which had the potential to help decision makers understand the potential progression of COVID-19 in their area. One such approach was developed by Kavadi et al. [132] who suggested a partial derivative regression and non-linear machine learning (PDR-NML) model to predict the COVID-19 transmission dynamics in India. The presented model achieved better performance in terms of accuracy and prediction time over linear regression and state-of-the-art AI-based models. It secured an accuracy of 99.7% with a prediction time of 5750 ms for 5000 samples. Another approach was offered by Carrillo-Larco and Castillo-Cara [133] who presented a model based on k-means clustering that categorized countries sharing similar numbers of confirmed SARS-CoV-2 cases. In this study, researchers not only used COVID-19 prevalence data (deaths, confirmed cases etc.), but also considered open access predictors like the prevalence of diseases such as HIV/AIDS, diabetes, and tuberculosis in 156 countries along with their standard health system metrics, air quality, and social-economic parameters such as the gross domestic production.

A third approach was developed by Hu et al. [134] , who attempted to develop an AI-based modified auto-encoder (MAE) that modeled multiple public health interventions by using real data to forecast a potential COVID-19 pandemic outbreak in a large geographic region and subsequently calculating the potential impact of interventions to curb the pandemic. The proposed architecture consisted of two single auto-encoders, each having three feedforward neural network layers. The used data pertained to confirmed positive COVID-19 cases and COVID-19 attributed deaths across 152 countries over the period of January 20 th to March 16 th , 2020. The estimate includes pandemic peak time, end time, and future cases. It concludes that MAE performed better in forecasting daily new cases in China (with an error of 0.00134) when compared with susceptible-exposedinfected-recovered (SEIR) model as used by Sameni [135] .

Beyond conventional ML-based approaches, advanced DL techniques have also been used to estimate future cases. Paul et al.

[136] suggested a convolutional LSTM-based multivariate spatiotemporal model to forecast the pandemic outbreak at the world-level. The proposed framework converted the spatial features into groups of temporal/non-temporal component-based 2D

images and used data from USA and Italy to train the network. The model performed well for predicting the number of cases over a 5-days period, with a mean absolute percent error (MAPE) of 0.3% and 5.57% for Italy and USA respectively. Other identified approaches for the development of potential COVID-19 medications and vaccinations are highlighted in Table VIII .

The investigation of this paper reveals several AI-based approaches that have been proposed as potential ways to help, with the COVID-19 pandemic, covering everything from initial diagnoses via image diagnostics up to the presentation of models that help to understand the spread of COVID-19 and identify potential new outbreak areas. By providing this comprehensive overview and identification of the current state-of-the-art, this structured review may help assist medical and research stakeholders currently involved in the COVID-19 response.

When it comes to potential barriers to the use of AI for COVID-19 responses, one of the foremost inaccuracies is related to the failure of forming diagnostic programs that fail differentiate by symptomatic basis. Therefore, in order to attain more accurate and precise prediction models, the patient's history of other existing ailments such as diabetes, heart, liver or other chronic conditions along with his age and gender should also be taken into account. The significance of these factors greatly influences the characteristics of COVID-19 for a specific patient compared with other pneumonia diseases known so far. As many countries around the world are opening up their data related to COVID-19, it becomes possible for anyone to create new AI-based services, and, additionally, if these statistics are provided in an effective way, tools for the prediction of COVID-19

diagnoses and spread could, in theory, be developed.

Another issue regarding the COVID-19 pandemic is urgency. As the government needs high-quality statistical data in almost real-time to understand the current epidemiological situation in their area and to regulate quarantine in that same area, effort should be focused on improving the availability and provision of data first. As at the beginning of the pandemic, data was often not of high quality (though this depends on a large amount on which country's data is being examined) the initially used scientific datasets were often constructed in a quick manner and may not be as precise or adequate as one would ideally like.

While considering the statistical analysis of the COVID-19 disease, the crucial challenge is to secure real-time high-quality data. As this disease is still in a rapid growth phase, the fast nature of data makes it extremely difficult to produce reliable models. Collection of regularly updated region-specific values of COVID-19 is crucial in order to overcome this issue.

A final challenge, and one of the most important, is the lack of understanding of epidemiological theory, beliefs, and knowhow of many AI researchers. This leads to a situation where many of the AI-based systems are created without input from medical professionals, thus limiting their usefulness and reliability. Thus, to this end, it is important that any implemented AIbased system aims to involve medical experts in the design and implementation.

Whilst it does appear to be the case that many AI-based approaches have been proposed, there is a limited real-world implementation of these tools. What this paper has shown, however, is that there is clearly a wide variety of potential methods that could be applied. Moving forward, it is imperative for AI-designers and researchers to work together with medical professionals to create and develop these systems that are applicable to real-world datasets and relevant to the work of the given stakeholder group (e.g. policymakers or health care workers). AI is not a cure-all that can fix the COVID-19 crisis, but it does have the potential to augment a given person's ability to handle and manage the crisis, thus, potentially, limiting the spread of the virus and leading to potentially better outcomes for infected patients.

As COVID-19 has permeated throughout the world, there is no one who has not felt its impact in some way. Thus, scientific and medical researchers are working rapidly to find a viable cure against the disease. One tool that may well help aid this effort is artificial intelligence (AI) by assisting front-line workers in numerous ways.

This paper has initially presented the details of the COVID-19 viruses including taxonomy, signs and symptoms, cause of spread, affected body organ, and hazardous impact on human society over the last century. Secondly, we mentioned the structural genome and origin of COVID-19 infectious virus, compared it with other Coroviridae family viruses, its transmission behavior and impact on global health. We presented an overview of potential use cases and AI-based implementations currently being developed and trialed. For instance, technological experts have developed AI-based programs for the rapid analysis, scanning and diagnosis of the coronavirus through pneumonia radiographic images or clinical blood sample data. These approaches offer an additional way to detect COVID-19, in addition to the more commonly used PCR test.

As testing capabilities continue to improve, AI-based tools become more useful to assist in clinical settings rather than to replace medical professionals. However, where AI is likely to see an increased use and benefit moving forward in the future, especially as more and higher quality data becomes available, is its use in predicting potential new areas of outbreak and forecasting the spread of COVID-19. These models are based on parameters such as the number of positive patients in different regions of a city and tracking their movements to anticipate the potential spread through contact. For example, it could be possible to apply an AI-based model to data gathered via a contact tracing application or sewer COVID-19 prevalence data, though other data could also be used, to better predict the scale and size of a new outbreak. The results thus obtained could help to manage and implement precautionary actions by defining guidelines in pandemic affected areas. It is important that AI-based tools are relevant, accurate, and impactful. In order to do this, cooperation is needed between AI-based researchers, medical professionals, and governmental agencies. Thus, further research should identify real-world empirical examples of AI being used on real-world data. This paper has made initial steps in gathering and highlighting the current state-of-the-art, but it does not differentiate between examples working -in the wild‖ and those that are done under laboratory conditions. So, while in theory, these approaches may well be useful, until they are evaluated within an applied context it is hard to quantify their level of usefulness, in fact, it may well be the case that AI-based tools actually inhibit effectiveness if they discreetly contain some level of bias, or replace the decision making capability of a more competent human actor.

☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. ☐The authors declare the following financial interests/personal relationships which may be considered as potential 

Origin and evolution of pathogenic coronaviruses

COVID-19: animals, veterinary and zoonotic links

Zoonoses. In: Encyclopedia of Food and Health

Defining epidemics in computer simulation models: How do definitions influence conclusions?

Defining epidemics in computer simulation models: How do definitions influence conclusions?

Smallpox -the death of a disease

Reassessing the Global Mortality Burden of the 1918 Influenza Pandemic

World Health Organization: 10 facts on Rabies

Hantavirus infections

World Health Organization: Dengue and severe dengue

World Health Organization: Marburg virus disease

Rotavirus infection

Ebola virus disease

Isolation of a Novel Coronavirus from a Man with Pneumonia in Saudi Arabia

COVID-19 Maps -John Hopkins Coronavirus Research Center

The species Severe acute respiratory syndrome-related coronavirus: classifying 2019-nCoV and naming it SARS-CoV-2

Isolation of a Novel Coronavirus from a Man with Pneumonia in Saudi Arabia

Seasonality of Respiratory Viral Infections

The epidemiology and pathogenesis of coronavirus disease (COVID-19) outbreak

Clinical course and risk factors for mortality of adult inpatients with COVID -19 in Wuhan, China: a retrospective cohort study. The Lancet

Detection of 2019 novel coronavirus (2019-nCoV) by real-time RT-PCR

Sensitivity of Chest CT for COVID-19: Comparison to RT-PCR

Correlation of Chest CT and RT-PCR Testing in Coronavirus 41

COVID-19 identification in chest X-ray images on flat and hierarchical classification scenarios

An automated Residual Exemplar Local Binary Pattern and iterative ReliefF based corona detection method using lung X-ray image. Chemometrics and Intelligent Laboratory Systems

Computed Tomography COVIDiag : a clinical CAD system to diagnose COVID-19 pneumonia based on CT findings

Machine Learning Analysis of Chest CT Scan Images as a Complementary Digital Test of Coronavirus

Coronavirus (COVID-19) Classification using Deep Features Fusion and Ranking Technique

Coronavirus (COVID-19) Classification using CT Images by

Explainable Deep Learning for Pulmonary Disease and Coronavirus COVID -19 Detection from X-rays

Covid-19: automatic detection from X-ray images utilizing transfer learning with convolutional neural networks. Physical and Engineering Sciences in Medicine

Classification of the COVID-19 infected patients using DenseNet201 based deep transfer learning

A Deep Learning System to Screen Novel Coronavirus Disease

Drawing insights from COVID-19-infected patients using CT scan images and machine learning techniques: a study on 200 patients. Environmental Science and Pollution Research

Development and Evaluation of an AI System for COVID-19 Diagnosis

A Fully Automatic Deep Learning System for COVID-19 Diagnostic and Prognostic Analysis

COVID-CT-Dataset: A CT Scan Dataset about COVID-19

Any unique image biomarkers associated with COVID-19?

Artificial intelligence-enabled rapid diagnosis of patients with COVID-19

Deep Transfer Learning Based Classification Model for COVID -19 Disease

Clinically Applicable AI System for Accurate Diagnosis, Quantitative Measurements, and Prognosis of COVID-19 Pneumonia Using Computed Tomography

Multi-task Deep Learning Based CT Imaging Analysis For COVID-19: Classification and Segmentation

AI Augmentation of Radiologist Performance in Distinguishing COVID-19 from Pneumonia of

Deep learning system to screen coronavirus disease 2019 pneumonia

Deep Learning Model for Diagnosis of Corona Virus Disease from CT Images

Classification of COVID-19 patients from chest CT images using multi-objective differential evolutionbased convolutional neural networks. European journal of clinical microbiology & infectious diseases : official publication of European Society of Clinical Microbiology

A deep learning algorithm using CT images to screen for corona virus disease (COVID -19)

A Novel Approach of CT Images Feature Analysis and Prediction to Screen for Corona Virus Disease (COVID-19)

Deep Learning-Based Recognizing COVID-19 and other Common Infectious Diseases of the Lung by

AI-assisted CT imaging analysis for COVID-19 screening: Building and deploying a medical AI

Deep Learning-based Detection for COVID-19 from Chest CT using Weak Label

Deep learning Enables Accurate Diagnosis of Novel Coronavirus

Convolutional capsnet: A novel artificial neural network approach to detect COVID -19 disease from X-ray images using capsule networks

Recognition of COVID-19 disease from X-ray images by hybrid model consisting of 2D curvelet transform, chaotic salp swarm algorithm and deep learning technique

Automated Deep Transfer Learning-Based Approach for Detection

Viral Pneumonia Screening on Chest Xray Images Using Confidence-Aware Anomaly Detection

Truncated inception net: COVID-19 outbreak screening using chest X-rays. Physical and Engineering Sciences in Medicine

Classification of COVID-19 from Chest X-ray images using Deep Convolutional Neural Networks

Can AI help in screening Viral and COVID-19 pneumonia? XX

CoroNet: A deep neural network for detection and diagnosis of COVID-19 from chest x-ray images

CovXNet: A multi-dilation convolutional neural network for automatic COVID-19 and other pneumonia detection from chest X-ray images with transferable multi-receptive feature optimization

Iteratively Pruned Deep Learning Ensemb les for COVID-19 Detection in Chest X-rays

Extracting Possibly Representative COVID -19 Biomarkers from X-ray Images with Deep Learning Approach and Image Data Related to Pulmonary Diseases

A modified deep convolutional neural network for detecting COVID-19 and pneumonia from chest X-ray images based on the concatenation of Xception and ResNet50V2

Weakly Labeled Data Augmentation for Deep Learning: A Study on COVID-19 Detection in Chest X-Rays

Interpretable artificial intelligence framework for COVID-19 screening on chest X-rays. Experimental and Therapeutic Medicine

COVID-Net: A Tailored Deep Convolutional Neural Network Design for Detection of COVID-19 Cases from Chest X-Ray Images

COVID-CAPS: A Capsule Network-based Framework for Identification of COVID-19 cases from X-ray Images

Finding Covid-19 from Chest X-rays using Deep Learning on a Small Dataset

Deep Learning on Chest X-ray Images to Detect and Evaluate Pneumonia Cases at the Era of COVID-19

A web-based Diagnostic Tool for COVID-19 Using Machine Learning on Chest Radiographs (CXR)

Detection of Coronavirus (COVID-19) Associated Pneumonia based on Generative Adversarial Networks and a Fine-Tuned Deep Transfer Learning Model using Chest X-ray Dataset

COVID-MobileXpert: On-Device COVID-19 Screening using Snapshots of Chest X-Ray

Within the lack of chest COVID-19 X-ray dataset: A novel detection model based on GAN and deep transfer learning

Shallow Convolutional Neural Network for COVID -19 Outbreak Screening using Chest

Automated detection of COVID -19 cases using deep neural networks with X-ray images

COVIDiagnosis-Net: Deep Bayes-SqueezeNet based diagnosis of the coronavirus disease 2019 (COVID-19) from X-ray images

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

COVID-ResNet: A Deep Learning Framework for Screening of COVID19 from Radiographs

Estimating Uncertainty and Interpretability in Deep Learning for Coronavirus (COVID -19) Detection

COVIDX-Net: A Framework of Deep Learning Classifiers to Diagnose COVID-19 in X-Ray Images

Automatic Detection of Coronavirus Disease (COVID-19) Using X-ray Images and Deep Convolutional Neural Networks

Deploying Machine and Deep Learning Models for Efficient Data-Augmented Detection of COVID-19 Infections

Automated diagnosis of COVID-19 with limited posteroanterior chest X-ray images using fine-tuned deep neural networks

Implementation of a Deep Learning-Based Computer-Aided Detection System for the Interpretation of Chest Radiographys in Patients Suspected for COVID-19

Using X-ray images and deep learning for automated detection of coronavirus disease

COVID_MTNet: COVID-19 Detection with Multi-Task Deep Learning Approaches

Improving Coronavirus ( COVID-19 ) Diagnosis using Deep Transfer Learning

Diagnosing COVID-19 Pneumonia from X-Ray and CT Images using Deep Learning and Transfer Learning Algorithms

COVID-19 diagnosis prediction in emergency care patients: a machine learning approach

Detection of COVID -19 Infection from Routine Blood Exams with Machine Learning: A Feasibility Study

Routine laboratory blood tests predict SARS-CoV-2 infection using machine learning

Machine learning ba sed approaches for detecting COVID-19 using clinical text data

Rapid and accurate identification of COVID-19 infection through machine learning based on clinical available blood test results

Abnormal respiratory patterns classifier may contribute to large-scale screening of people infected with COVID-19 in an accurate and unobtrusive manner

Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation

AI4COVID-19: AI enabled preliminary diagnosis for COVID-19 from cough samples via an app

Exploring Automatic Diagnosis of COVID-19 from Crowdsourced Respiratory Sound Data

Detection of Respiratory Infections using RGB -infrared sensors on Portable Device

Predicting COVID-19 Malignant Progression with AI Techniques

Helping doctors hasten COVID-19 treatment: Towards a rescue framework for the transfusion of best convalescent plasma to the most critical patients based on biological requirements via ml and novel MCDM methods

Early prediction of mortality risk among severe COVID-19 patients using machine learning

Deep transfer learning artificial intelligence accurately stages COVID-19 lung disease severity on portable chest radiographs

Association between weather data and COVID -19 pandemic predicting mortality rate: Machine learning approaches

COVID -19 Patient Health Prediction Using Boosted Random Forest Algorithm. Frontiers in Public Health

Evaluation of novel coronavirus disease (COVID-19) using quantitative lung CT and clinical data: prediction of short-term outcome

Artificial Intelligence Approach to Predict the COVID -19

Patient ' s Recovery Artificial Intelligence Approach to Predict the COVID-19 Patient ' s Recovery

Severit y Detection For the Coronavirus Disease 2019 (COVID-19) Patients Using a Machine Learning Model Based on the Blood and Urine Tests

Machine learning-based CT radiomics model for predicting hospital stay in patients with pneumonia associated with SARS-CoV-2 infection: A multicenter study

A Machine Learning Model Reveals Older Age and Delayed Hospitalization as Predictors of Mortality in Patients with COVID-19

Severity Assessment of Coronavirus Disease 2019 (COVID -19) Using Quantitative Features from Chest CT Images

A machine learning-based model for survival prediction in patients with severe COVID-19 infection

Partial Derivative Nonlinear Global Pandemic Machine Learning Prediction Of Covid 19

Using country-level variables to classify countries according to the number of confirmed COVID-19 cases: An unsupervised machine learning approach

Forecasting and Evaluating Multiple Interventions for COVID-19 Worldwide

Mathematical Modeling of Epidemic Diseases; A Case Study of the COVID-19 Coronavirus

A multivariate spatiotemporal spread model of COVID-19 using ensemble of ConvLSTM networks

Study of ARIMA and least square support vector machine (LS -SVM) models for the prediction of SARS-CoV-2 confirmed cases in the most affected countries

Deep learning methods for forecasting COVID-19 time-Series data: A Comparative study

Prediction of epidemic trends in COVID-19 with logistic model and machine learning technics

A Novel Adaptive Deep Learning Model of Covid-19 with focus on mortality reduction strategies

Analysis on novel coronavirus (COVID-19) using machine learning methods

Analysis of Spatial Spread Relationships of Coronavirus (COVID-19) Pandemic in the World using Self Organizing Maps. Chaos, Solitons and Fractals

Predicting the growth and trend of COVID-19 pandemic using machine learning and cloud computing

Predicting COVID -19 Incidence Through Analysis of Google Trends Data in Iran: Data Mining and Deep Learning Pilot Study. JMIR Public Health and Surveillance

Real-time forecasts and risk assessment of novel coronavirus (COVID-19) cases: A data-driven analysis

Neural Network aided quarantine control model estimation of COVID spread in Wuhan

Composite Monte Carlo decision making under high uncertainty of novel coronavirus epidemic using hybridized deep learning and fuzzy rule induction

SEIR and Regression Model based COVID-19 outbreak predictions in India

Generating Similarity Map for COVID-19 Transmission Dynamics with Topological Autoencoder

Variational-LSTM Autoencoder to forecast the spread of coronavirus across the globe

COVID-19 in Iran: A Deeper Look Into The Future

Predicting the epidemic curve of the coronavirus (SARS-CoV-2) disease (COVID-19) using artificial intelligence

A machine learning methodology for realtime forecasting of the 2019-2020 COVID-19 outbreak using Internet searches, news alerts, and estimates from mechanistic models

Prediction for the spread of COVID-19 in India and effectiveness of preventive measures

Forecasting the Worldwide Spread of COVID-19 based on Logistic Model and SEIR Model

Transmission dynamics of the COVID-19 outbreak and effectiveness of government interventions: A data-driven analysis

Multiple-Input Deep Convolutional Neural Network Model for COVID-19 Forecasting in China

Modified SEIR and AI predi ction of the epidemics trend of COVID-19 in China under public health interventions

Vaxign-ML: supervised machine learning reverse vaccinology model for improved prediction of bacterial protective antigens

COVID-19 Coronavirus Vaccine Design Using Reverse Vaccinology and Machine Learning

XGBoost: A Scalable Tree Boosting System

Potential Neutralizing Antibodies Discovered for Novel Corona Virus Using Machine Learning

Deep learning-based computational drug discovery to inhibit the RNA Dependent RNA Polymerase: application to SARS-CoV and COVID-19

Protease Inhibitors Designed Using Generative Deep Learning Approaches and Reviewed by Human Medicinal Chemist in Virtual Reality

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

CRISPR-based surveillance for COVID-19 using genomically-comprehensive machine learning design

Potential 2019-nCoV 3C-like protease inhibitors designed using generative deep learning approaches DNA double-strand break repair in mammalian cells. View project Matrix-isolated systems modeling View project

competing interests: