Machine Learning Research Towards Combating COVID-19: Virus Detection, Spread Prevention, and Medical Assistance

Osama Shahid∗, Mohammad Nasajpour∗, Seyedamin Pouriyeh∗, Reza M. Parizi†, Meng Han∗, Maria Valero∗, Fangyu Li ‡, Mohammed Aledhari§, Quan Z. Sheng¶ ∗ Department of Information Technology, Kennesaw State University, Marietta, GA, USA {oshahid1, mnasajp1}@students.kennesaw.edu, {spouriye, mhan9, mvalero2}@kennesaw.edu † Department of Software Engineering and Game Development, Kennesaw State University, Marietta, GA, USA [email protected] ‡Department of Electrical and Computer Engineering, Kennesaw State University, Marietta, GA, USA fl[email protected] §Department of Computer Science, Kennesaw State University, Marietta, GA, USA [email protected] ¶Department of Computing, Macquarie University, Sydney, Australia [email protected]

Abstract—COVID-19 was first discovered in December 2019 normal carrying the virus and continuing to spread the disease and has continued to rapidly spread across countries worldwide without knowing [4]. As COVID-19 has a rapid nature of infecting thousands and millions of people. The virus is deadly, spreading, the World Health Organization (WHO) declared and people who are suffering from prior illnesses or are older than the age of 60 are at a higher risk of mortality. Medicine and it as a global pandemic in March 2020 [5]. At the time of Healthcare industries have surged towards finding a cure, and writing this paper (i.e., September 2020), the total number of different policies have been amended to mitigate the spread of confirmed COVID-19 cases worldwide was over 32 million the virus. While Machine Learning (ML) methods have been [6]. To tackle this outbreak, scientists in different research widely used in other domains, there is now a high demand communities are seeking a wide variety of computer-aided for ML-aided diagnosis systems for screening, tracking, and predicting the spread of COVID-19 and finding a cure against systems such as the Internet of Things [7], Machine Learning it. In this paper, we present a journey of what role ML has (ML) or Deep Learning (DL) techniques [8], [9], Big Data played so far in combating the virus, mainly looking at it from [10], and Blockchain [11] that can assist with overcoming the a screening, forecasting, and vaccine perspectives. We present a challenges brought by COVID-19. These technologies can be comprehensive survey of the ML algorithms and models that can used for controlling the spread of the virus, detecting the virus, be used on this expedition and aid with battling the virus. Index Terms—COVID-19, Machine Learning, Artificial Intel- or even designing and manufacturing a vaccine or drug to ligence, Healthcare, Drug Development, Prevention, Predictive combat it. Analysis, Diagnosis, Image Classification. There were two epidemics in the past from the coronavirus family including Severe Acute Respiratory Syndrome (SARS- I.INTRODUCTION CoV) [12] and Middle Eastern Respiratory Syndrome (MERS) arXiv:2010.07036v1 [cs.CY] 29 Sep 2020 In December 2019, a novel severe contagious respiratory [13]. SARS-CoV is a respiratory virus that was transmissible syndrome coronavirus 2, which is a type of Severe Acute from person to person and it was first identified in 2003. The Respiratory Syndrome (SARS-CoV-2) virus called COVID-19, virus had over 8,000 confirmed cases worldwide during its was discovered in Wuhan, China [1]. COVID-19 virus is course which affected over 26 countries [14]. MERS is also a airborne and can easily spread and infect people [2]. Ac- respiratory virus with similar symptoms of SARS-CoV. cording to the Centers for Disease Control and Prevention ML, as a subset of Artificial Intelligence (AI), has shown a (CDC) [3], the infected people show a range of symptoms lot of potentials in many industries like retail [15], banks [16], like dry cough, shortness of breath, fatigue, losing the sense of healthcare [17], [18], pharmaceuticals [19], and many more taste and smell, diarrhea, and congestion. Infected patients can [20]. ML techniques can be programmed to imitate human also present fever episodes. Strangely enough, some patients intelligence. For example, in the healthcare industry, ML who have contracted the virus might not even show any of techniques can be trained and used towards medical diagnosis the aforementioned symptoms [4]. They can feel completely [21]. ML models have been vastly trained over a dataset consisting of medical images like Computed Tomography (CT) Corresponding author: Seyedamin Pouriyeh (email: [email protected]). Scan, Magnetic Resonance Imaging (MRI), or X-Ray to detect anomalies [22], [23]. Its classification models can be expanded process of treatment. Regarding the COVID-19, not only it in diverse areas including cancer [24], diabetes [25], fatty liver helps to avoid the spread of contamination, but also it is cost- [26], etc. As an example, breast cancer can be diagnosed with beneficial. The standard method of diagnosing COVID-19 is a prediction accuracy of 97.13% [27] using ML models. to conduct Reverse-Transcription Polymerase Chain Reaction During previous epidemics, ML techniques have been (RT-PCR) test [36]. The RT-PCR is a swab test that is used widely implemented in order to assist healthcare authorities to detect nucleic acid from COVID-19 in the upper and for better actions regarding the diseases [28]. For example, lower respiratory system. At the beginning of the pandemic, Sandhu et al. [29] proposed an ML model that utilizes GPS the sensitivity of the RT-PCR test could show negative for technology along with cloud computing power and Google patients who were later in-fact confirmed positive, hence the Maps to represent potentially infected patients and provide an reporting of false-negative rates was high [37]. There was also alternative route for uninfected users resulting in potentially a concern about having a shortage and a limited number of mitigating the spread. The model reaches the classification tests, and the high-cost factor of producing and conducting accuracy of 80% in re-routing away from infected patients. In them [38]. It is important to explore alternative methods another study, Choi et al. [30] used ML models for sentimental of diagnosing COVID-19 that would speed-up the process analysis to review public overreaction appearing in media [39]. To help overcome the challenge presented in diagnosing articles and social media platforms. This type of in-depth COVID-19. ML models and algorithms have shown promising analysis can rapidly monitor the public reaction. It can also results in different stages of COVID-19 [40]. In general, ML aid policymakers in taking the right actions in reducing fear techniques have been widely utilized in the healthcare domain and distress from the public regarding MERS. and similarly, it can be used towards analyzing data and ML has also been widely used in order to improve clinical diagnosing COVID-19 using medical imaging which includes decision-making regarding the current COVID-19 pandemic X-Ray and CT Scan images [34]. In this section, we review [31]. Researchers, using ML algorithms and clustering tech- different ML techniques that have been used for screening nique, are able to forecast the spread in provinces [32], [33]. and diagnosis of COVID-19. Moreover, we discuss other ML- ML methods of image classification are used by the scientific based tools including Chatbots and Artificial Intelligence of community to help in diagnosing the deadly virus [34]. With Things (AIoT). the objective of finding a cure for the virus, ML algorithms are used to evaluate how dependable are off-the-counter drugs A. Medical Imaging may be used to help infected patients [35]. The aforementioned examples show the potential of ML in Diagnosing COVID-19 is one of the most important parts the detection, diagnosis and prediction of viruses. This paper of dealing with the disease. As a result of low access and largely provides a survey reviewing the research that has been high possibility of false-negative results to the RT-PCR kits, dedicated by the scientific community in using ML technology there is an essential need for using other approaches such as in combating COVID-19. In particular, we investigate the role medical images analysis for accurate and reliable screening of ML in detecting or screening, forecasting, and medical and diagnosis in COVID-19 [41]. In general, analyzing medi- assistance for the virus. cal imaging modalities such as chest X-ray and CT-Scan have The remainder of the paper is organized as follows. In Sec- key contributions in confirming the diagnosis of COVID-19 as tion II, we review the role of ML in detecting and the screening well as screening the progression of the disease [42]. Different process of COVID-19. We study the use of different ML ML techniques that incorporate X-ray and CT-Scan image techniques regarding four major sections for diagnosing and processing approaches could help physicians and healthcare screening including Medical Imaging, Chatbot, and Artificial professionals as a better way for diagnosis and understanding Intelligence of Things (AIoT). In Section III, the importance of of the progression of the COVID-19 disease. disease contamination and its exposure to others are discussed 1) X-ray: During this pandemic, chest imaging can be an regarding the use of ML for tracking and predicting the important part of the COVID-19 in early stage of detection. spread of COVID-19. This section is mainly divided into three Classifying patients rapidly is what is expected from these parts reviewing preventing the spread, contact tracing, and approaches. Within the categorization of medical imaging, forecasting. Similarly, Section IV reviews the need for medical Chest X-Ray (CXR) was recommended to be implemented as assistance during the pandemic and how ML technology can the first medical imaging regarding COVID-19 by the Italian be integrated. Understanding the virus and developing a drug Society of Radiology (SIRM) [43]. Figure 1 demonstrates the or vaccine using ML techniques are discussed in this section CXR images from infected and normal people. According as well. Finally, we discuss, outline future work, and conclude to Cozzi et al. [44], CXR has a sensitivity of 67.1% which in sections V and VI respectively. can be first implemented in special cases including assisting radiologists with better COVID-19 cases identification and II.MLTECHNIQUES TOWARDS DETECTING &SCREENING fast treatment assigning to the patient. Additionally, CXR is COVID-19 inexpensive and secure because of minimizing the risk of Detecting either a symptomatic or asymptomatic disease in contamination which makes a safer workplace for healthcare an early stage could be highly effective in order to start the workers as well. In order to decrease the amount of work by radiologists, ML 2) CT-Scan: Another applicable medical imaging tool for techniques can be assigned to classify patients with respect COVID-19 diagnosis is a chest Computed Tomography (CT) to COVID-19. To do that, researchers are mostly focused on Scan which is more accurate in detecting COVID-19 cases the ML classification models such as Support Vector Machine [79]. Due to respiratory problems of COVID-19 which include (SVM), Convolutional Neural Networks (CNN), DL. One lung abnormalities, CT-Scan can be specified as the detecting approach [45] implemented X-Ray in order to classify the lung procedure for the early stage of a pandemic while none of lesions (caused by COVID-19) with Multi-level Threshold the COVID-19 symptoms appear in patients [80]. Figure 2 (MT) process and SVM model. Within this model, firstly, demonstrates the CT-Scan images from infected and normal the lung image contrast will be enhanced. Secondly, the people. image will be reduced into specific sections (using MT) to Ardakani et al. [81] implemented a Computer-aided diagno- avoid duplication of work on uninfected areas. Lastly, the sis system to show the benefits of DL in diagnosing COVID-19 SVM model classifies the sections of the lung with respect using a variety of CNNs which conclude the ResNet-101 as to the predefined healthy lungs. Sethy et al. [46] developed the most precise model. The proposed model can lower the a platform using a variety of Deep Convolutional Neural workload of radiology workers as well. Similarly, Li et al. [82] Networks (DCNN) models classifying within the SVM with developed a DL framework known as COVID-19 detection two different datasets in order to detect COVID-19 cases based neural network (COVNet) which can differentiate COVID-19 on the related CRX image. Similarly, [47] proposed a DCNN from typical types of pneumonia using chest CT-Scan images. model using the data gathered from two hospitals in Italy to An ML model for quantitative infection assessment through represents the importance of AI in the detection of COVID-19. CT Scan images was modeled by Shan et al. [83] suggesting their model is capable of estimating the shape, volume, and the Zhang et al. [48] trained ML models over a large viral percentage of infection. A Human-In-The-Loop method [84] pneumonia dataset of CXR images to detect anomalies. They was proposed that involves healthcare workers to intervene tested their model on a completely different dataset that has with the VB-Net (a modified 3D CNN that combines V-Net COVID-19 CXR images. This is done as one of the symptoms and bottle-neck structure) ML model to add more CT-Scan of COVID-19 can be pneumonia [3]. The results are impres- images into the training model to constantly update the model sive as the model performs well when tested on the COVID-19 and produce more efficient results. Tang, et al. [85] introduced dataset with the Area Under the Curve (AUC) of 83.61%. It is another method to detect the severity of COVID-19 through even more impressive as the model was trained on a different assessing quantitative features from CT-Scan images. The use dataset and yet performed well. Similarly, Wang et al. [49] of the Random Forest (RF) model including 500 decision utilized COVIDx dataset, a publicly available dataset consists trees along with three-fold cross-validation allows authors to of COVID-19, pneumonia and non-COVID-19 pneumonia- calculate 63 quantitative features of COVID-19 like infection related X-ray images. The authors used this data to train their volume or lung ratio. However, the model is limited by a model for detection of COVID-19, the Deep Neural Network binary classification i.e., results can either be severe or non- (DNN) is referred to as COVID-Net showing promising results severe when they should be mild, common, severe, and critical. in diagnosing infected patients. Apostolopoulos et al. [50] Gozes et al. [86] trained clinical models integrated with ML used transfer learning approaches like feature extraction and to detect the virus that achieves high accuracy and is also able fine-tuning of CNN based models and trained and tested over to quantify the burden of the disease. similar datasets achieving a prediction accuracy up to almost Similar to the previous section, ML techniques that can 98%. They demonstrated that implementing transfer learning be used for detecting COVID-19 from CT-Scan images are can have a significant improvement in results. Most ML classi- presented in Table II. The list of references includes both fiers are trained and tested to achieve high prediction accuracy published and yet to be peer-reviewed. of COVID-19; however, it is also important to quantify the uncertainty that could exist by using such classifiers as a B. Chatbots primary medium of diagnosis. An approach to validate the Computer programs developed to communicate with hu- ML prediction of diagnosis in CXR images was reviewed by mans by adopting natural languages are called chatbots [103]. Ghoshal et al. [51]. It exploited a Bayesian Deep Learning A chatbot basically can communicate with different users classifier to estimate the model uncertainty. The result analysis and generate proper responses to those users based on their displays a strong correlation between uncertainty and accuracy inputs. Recently, the COVID-19 pandemic has led to building of prediction, which means that the higher the uncertainty different chatbots instead of using hotlines as a communication outcome, the more reliable the prediction accuracy. method. This will reduce hospital visits and increase the Many other ML models are utilized for detecting efficiency of communicating [104]. Generally, chatbots are COVID-19, and a subset of them are presented in Table I. implemented in order to provide an online conversation with The references are both published papers and also papers that the user by either text or voice displays on web applications, are yet to be peer-reviewed. All references that are presented in smartphone applications, channels, and so on [105]. This the table show high performance towards detecting predicting conversation can help the user to have a better understanding COVID-19 through CXR images. of his or her situation and gives some hints to users so Fig. 1. Chest X-Ray (CXR) images of COVID-19 infected people versus uninfected people [52].

TABLE I MLRESEARCHDONETOWARDSDIAGNOSING COVID-19 USING X-RAY (CXR) DATASETS.

Reference Dataset Methods Remarks [50] Multiple Datasets that include 448 confirmed COVID-19 images source: Github DL - CNN, Feature Extraction (various models) Various model performance comparison

[51] 68 COVID-19 Chest X-ray images and 5873 Pneumonia images source: Github, Kaggle Bayesian Deep Learning Classifiers Using Transfer learning approach with the classifier to estimate uncertainty

[53] Fine Tuning - ResNet Model Model achieves high accuracy for multi-class classification [54] Multiple models VGG19, MobileNet Test and train multiple image classifiers to obtain the highest accuracy identifying the virus; VGG19 and DenseNet have a high accuracy score [55] Multiple Image Classification Models Demonstrate how DNNs are vulnerable to a universal adversarial perturbation causing failure in classification tasks, fine-tuning may be able to improve this [56] Capsule Network-based framework - COVID-CAPS An alternative modeling framework that has decent performance with low trainable parameters COVIDx [57] DenseNet-121 The model is explained well, to test its robustness authors perform multi-class classification and perform k-fold validation [58] SqueezeNet-Bayesian based model The authors classify input of X-Ray images of Normal, COVID-19, Pneumonia using data augmentation and fine-tuning techniques [59] CoroNet - Semi-supervised learning technique based on AutoEncoders [60] EfficientNet The proposed approach is able to produce a model with high quality with high accuracy [49] COVID-Net COVID-Net is publicly accessible for scientists to further build and improve the network to achieve high accuracy [61] CXR images of 50 COVID-19 patients, and 50 Normal CXR images Three different CNN models (InceptionV3, ResNet50, Inception-ResNetV2) ResNet50 provides the highest classification.

[62] 170 Chest X-Ray images of 45 patients from 5 different sources Modified Pre-trained AlexNet and a simple CNN Pre-Trained network achieves a higher accuracy

[63] 295 COVID-19 CXR images and 163 Pneumonia and Normal CXR images MobileNetV2/SqueezeNet They use DL models achieve high classification rate

[64] 127 COVID-19 CXR images and 1000 Pneumonia and Normal CXR images DarkCovidNet (CNN) Provides two approaches, a binary classification and multi-class classification. Binary model performs better and higher

[65] A large dataset containing both COVID-19 and Non COVID-19 cases from multiple sources VGG16, InceptionV3, Xception, DenseNet-121, NasNet-Mobile, etc models compared VGG16 achieved the highest rate

[66] 306 COVID-19 CXR images and 113 normal CXR images CNN model using Decision Tree Classifier The tested method appears to be robust and provides results that are accurate

[67] CXR images of confirmed 150 COVID-19 patients from Wuhan source: Kaggle Convolutional Neural Network The authors are able to get 93% accuracy

[46] Multiple datasets with 183 COVID-19 CXR images and CXR images of SARS-CoV and MERS Trained 9 different models for COVID-19 ResNet50 + SVM achieved the highest accuracy

[68] 231 COVID-19 CXR images, and 2100 Pneumonia and normal CXR images source: Github Novel ANN - Convolutional CapsNet Model provides highly accurate diagnostics with Binary Classification (whether COVID-19 or no finding)

[69] 423 COVID-19 CXR images, 3064 CXR images of normal CXR and viral pneumonia 8 Different CNN Models Transfer Learning and Data augmentation Techniques used, CheXNet performs the best

[70] 192 COVID-19 CXR images, and 145 normal CXR images nCOVnet(VGG-16) The proposed model is able to achieve high accuracy in predicting COVID-19 infected patients from CXR Images

[71] 10 CXRs from COVID-19 confirmed patients in China and USA DL U-Net Model Model shows great promise, with potential use towards early diagnosis for COVID-19 pneumonia

[72] 126 COVID-19 CXR images, 5835 normal and pneumonai CXR images GSA-DenseNet121-COVID-19 (Hybrid CNN using Optimization algorithm) The proposed model achieves high accuracy in diagnosing, up to 98%

[73] 250 COVID-19 CXR images, 4934 Non COVID-19 CXR images ResNet18, ResNet50, SqueezeNet and DenseNet-121 The models perform well and are tested across multiple parametrs such as Receiver Operating Characteristic (ROC), precision-recall curve, etc.

[74] Multiple datasets including 162 COVID-19 positive CXR images, and Non COVID-19 CXR mages Truncated Inception Net The proposed model achieves an accuracy of 99.92% (AUC 0.99) in classifying COVID-19 positive cases

[75] 305 COVID-19 CXR images, and 822 Non COVID-19 CXR images Transfer Learning method employed on pre-trained models The authors use Gradient Class Activation Map for detecting where the model focuses more on for classification

[76] 180 COVID-19 CXR images, and Non COVID-19 CXR images Xception and ResNet50V2 A concatenated of the two models performs well towards detecting COVID-19

[77] 318 COVID-19 CXR images, and Non COVID-19 CXR images COVID-DA Propose a Deep Learning model that has a novel classifier separation scheme

[78] 455 COVID-19 CXR images, and 3450 Non COVID-19 CXR images MobileNetV2 Training the CNN MobileNetV2 model from scratch proves to get higher accuracy compared to transfer learning techniques that he or she can take proper steps. Chatbots are usually Google Cloud Platform with the main assistance of Natural considered as one of the best suited to screen patients remotely Language Processing (NLP) which is compatible with either without interactions [106]. The advantages of them include speech or text. Similarly, Ouerhani et al. [109] developed a quickly updating information, repetitively encouraging new chatbot (called “COVID-Chatbot”) based on DL model which behaviors such as washing hands, and assisting with psy- uses NLP in order to enhance the awareness of people about chological support due to the stress caused by isolation and the ongoing pandemic. COVID-Chatbot is implemented to misinformation [107]. The ML-based chatbots are improved decrease the impact of the disease during and after the quar- during the training procedure and using more data makes antine phase. Another example is Bebot [110] that provides this approach more reliable. During the COVID-19 pandemic, updated data regarding the pandemic, and also assists patients chatbots are getting more attention in order to provide more with symptoms checking. Some other implemented chatbots details about COVID-19 in different stages. A wide variety during this unprecedented time are including Orbita [111], of chatbots with different languages have been implemented Hyro [112], Apple’s screening tools (website, application, to help patients at the early stage of COVID-19. “Aapka voice command or Siri) [113], CDC’s self-checker [114], and Chkitsak”, an AI-based chatbot developed by [108] in India, Symptoma [115]. A brief description of these chatbots can be assists patients with remote consultation regarding their health found in Table III. information, and treatments. This application is developed on Fig. 2. CT-Scan images of COVID-19 infected people versus uninfected people [52].

TABLE II MLRESEARCHDONETOWARDSDIAGNOSING COVID-19 USING CT-SCAN DATASETS.

Reference Dataset Methods Remarks [87] Open-sourced COVID-CT source: Github Multi-task and Self-Supervised learning Clinically useful

[85] Clinical CT scan images of 176 COVID-19 cases Random Forest / Three-fold cross-validation The Random Forest model showing promising performance for reflecting the severity of COVID-19

[86] Multiple International Datasets of CT scan images (Chinese CDC, Hospitals from China and USA, and Chainz.cn) ResNet50 High accuracy in identifying whether COVID-19 cases

[82] 4356 CT scan images (including COVID-19 and Non COVID-19 scans) collected from 6 Hospitals Present a Deep Learning model CovNET The model has the ability to high accuracy in identify COVID-19 cases from other lung diseases

[88] 618 Clinical CT scan images (including COVID-19 and Non COVID-19 scans) Deep Learning - ResNet18 Using a location attention mechanism on the model distinguishes COVID-19 cases from others with higher accuracy

[89] Clinical CT Scan Images from 99 Patients from 3 Hospitals (including COVID-19 and Non COVID-19 scans) Modified Transfer Learning - Inception Model Use a fine-tuning technique with pre-trained weights

[90] Clinical CT scan images from 133 Patients from Hospital in China Multi Stage - DL Models, LSTM The model is capable of extracting spatial and temporal information efficiently thereby improving prediction performance

[91] CT scan images of 413 COVID-19 cases and 439 of pneumonia or normal cases ResNet50 The model with transfer learning technique performs better than alternative supervised learning methods

[92] CT scan images obtained from 5 Hospitals in China Various DL learning models (Inception, ResNet50, 3D U-Net++) Able to deploy the model in 4 weeks overcoming challenges and achieving a high accuracy rate

[83] 549 CT Scan images obtained from clinics in China Deep Learning - VB-Net Introducing a Human-in the loop section to refine automation of each case - the model can segment and quantify infected regions

[93] Large-scale dataset including 10,250 CT scan images of COVID-19 and Non COVID-19 scans UNet - 2D Segmentation DL CNN Model The DL models outperforms radiologists in diagnostic performance

[94] Clinical CT scan images of 558 COVID-19 patients with pneumonia, collected from 10 hospitals COPLE-Net + Noise-Robust Dice Loss The technique presented outperforms standard Noise-Robust loss functions, COPLE-Net and the framework both perform quite highly in segmenting labels for COVID-19 pneumonia lesion

[95] Clinical CT scan images of 83 COVID-19 cases and 83 Non COVID-19 cases BigBiGAN (bi-directional generative adversarial network) The model achieves high validation accuracy in identifying COVID-19 pneumonia from CT images

[96] Large-scale dataset that include 400 COVID-19, and more Non COVID-19 scans Classification, Segmentation and Encoder-Decoder Model - Res2Net Model is highly efficient for Classification and Segmentation

[97] Multiple datasets that include 473 COVID-19 CT scans UNet Propose a method to incorporate spatial and channel attention

[98] Dataset of CT-Scans from 1,684 COVID-19 patients Inception V1 Validate the model in 3 ways including 10-fold cross-validation achieving high AUC for the validation dataset

[99] Clinical CT scan images including 146 COVID-19 cases and 149 Non COVID-19 cases DenseNet The model classifies COVID-19 over CT Images with high AUC

[34] 219 CT scan images of COVID-19 and 399 CT scan images of normal or other diseases VGG-16, GoogleNet, ResNet Support Vector Machine (SVM) was used for binary classification

[100] 746 CT scan images of COVID-19 and Non COVID-19 cases; Open-Sourced - Github Capsule Networks (CapsNets), ResNet Authors present a detail oriented capsule network, implementing data augmentation techniques to overcome lack of data

[101] Clinical CT scans of 88 confirmed patients confirmed with COVID-19 from hospitals in China DeepPneumonia (ResNet-50) Model is capable of predicting COVID-19 with high accuracy

[102] 1,129 Clinical CT scan images for COVID-19 detection UNet, Weakly Supervised DL Network (DeCovNet) Accurately predict COVID-19 infectious probability without annotating lesions for training

C. Artificial Intelligence of Things (AIoT) A similar device has been developed in Taiwan for a hospital In general, applications of the Internet of Things (IoT) with the collaboration of Microsoft in order to detect face [117]–[119] and AI can assist businesses with process automa- mask wearings and temperatures. Consequently, any detection tion which would decrease the contacts of humans due to the of proposed problems can easily be reported to the staff which lower number of people needed [120]. During the COVID-19 will maintain an uncontaminated atmosphere [122]. pandemic, AI and IoT are getting more attention in the healthcare domain where screening and detecting procedures can be done more safely. Thermal imaging and social distance monitoring are two main functions that are mainly considered in the screening phase of COVID-19. In fact, the aims of those devices are high-temperature detection, face mask screening, and distance controlling that are discussed in the coming sections. 1) Thermal Imaging: With respect to the IoT thermal screening applications, AI can assist in this area by implement- ing appropriate algorithms. SmartX [121], a thermal screening device using infrared thermal imaging and AI face recognition makes screening in crowd buildings or entrances more efficient (see Figure 3). The device captures a visitor’s temperature Fig. 3. An industrial thermal imaging system enabled with AI [121]. and also checks whether he or she is wearing a face mask. TABLE III AI CHATBOTS/VIRTUAL ASSISTANTS COMBATING COVID-19.

Reference AI chatbot/virtual assistant Name Origin Company Function [108] Aapka Chkitsak India Academic Research Remote consultation

[109] COVID-Chatbot Tunisia & Germany Academic Research Enhancing awareness regarding COVID-19

[110] Bebot Japan Bespoke Update information & Symptoms checker

[111] Orbita USA - Reduce contacts

[112] Hyro Israel - Interacting with patients

[116] Symptoma Austria - Diagnosing by symptoms checking

[116] COVID-BOT France Clevy.io Assist with symptoms by knowledge of government and WHO

2) Social Distance Monitoring: Regarding the necessity the spread of COVID-19. ML models can be utilized regarding of practicing social distancing using IoT devices, AI can the forecasting of the virus in providing early-signs of the implement an automated screening using computer vision COVID-19 and projecting its spread. Also, contact tracing and methods [123]. An instance of such a device is RayVision social media data analysis have shown promising results in [124] which ensures social distancing and face mask wearing COVID-19 spread mitigation [128]. The scope of this section guidelines are followed in the crowd. By using the computer is to provide a review of the research of ML tools and models vision techniques, it can monitor people with a live stream that can make this possible. on its specific dashboard which allows alerting the authorities in case of any rule-breaking [125]. Figure 4 represents the A. Early Signs, Preventing the Spread process of monitoring using cameras. Similarly, Landing AI Obtaining early-warning signs for an outbreak of an epi- [126] is another AI-based technology which can detect social demic could really help towards slowing and mitigating the distancing violations in real-time. Moreover, a peer-reviewed spread of the virus. It also can encourage societies to take research [127] implemented an Unmanned Aerial Vehicle necessary precautionary measures [129]. In this section, we (UAV) or a drone with the ML application in response to the review the early-warning signs that were made possible using need for maintaining social distance in crowds. Interestingly, the ML technology. The World Health Organization (WHO) masks will be provided by the drone for people who do not made statements about COVID-19 being a potential outbreak wear a mask. on the 9th of January 2020. There were AI companies like BlueDot and Metabiota that were able to predict the outbreak even earlier [130]. BlueDot focuses on spotting and predicting outbreaks of infectious diseases using its proprietary methods and tools. They use ML and NLP techniques to filter and focus the risk of spreading a virus. Using the data from local news reports of first few suspected cases of COVID-19, historical data on animal disease outbreaks, and airline ticket information, they were then able to use their tool to predict a definite outbreak occurring within nearing cities and other regions of China [33]. BlueDot had warned its clients about the outbreak on the 31st December 2019, over a week prior Fig. 4. An industrial screening system for monitoring the social distance of to the WHO made any statements about it [130]. Similarly, people and their personal protective equipment [124]. Metabiota used their ML algorithms and Big Data to predict outbreaks and spreads of diseases, and event severity [33]. They used their technology and flight data to predict that III.MLTECHNIQUESTOWARDSPREDICTINGAND there will be a COVID-19 outbreak in countries like Japan, TRACKINGTHESPREADOF COVID-19 Thailand, Taiwan, and South Korea. As we are currently in the midst of a global pandemic, the ability to predict and forecast the spread of the COVID-19 B. Contact Tracing could i) help the general population in taking preventative One of the major approaches for preventing the spread measures, ii) allow healthcare workers to anticipate and pre- of the virus is tracing the confirmed cases of COVID-19 pare for the wave of potentially infected patients, and iii) allow because of the potential spread of the virus through droplets policymakers to make better decisions regards the safety of the by coughs, sneeze, or talking [131]. It is recommended that general population. More importantly, the ability to predict the not only the people who have a positive COVID-19 test, spread of the virus can be used to mitigate and even prevent but also the ones who had been in close contact with the confirmed cases to be quarantined for 14 days. The contact using ML approaches for forecasting the spread of COVID-19 tracing applications are applied all over the world for this is getting lots of attention among the research communities. purpose with different methods. Basically, it starts after the Hu et al. [32] proposed an unsupervised ML method diagnosis process because the detected case needs to be for forecasting. They used a Modified Auto-Encoder model traced. Most importantly, after the data is collected by those (MAE) and trained it to predict the transmission of COVID-19 applications, ML and AI techniques will start analyzing for cases across 31 provinces or cities in China integrating K- discovering further spread of the disease [132]. Although the means algorithms to achieve a high prediction accuracy. Yang contact tracing applications could be deeply helpful during the et al. [159] used epidemiological data of COVID-19 in an pandemic, privacy issues can bring high concerns regarding the SEIR (Susceptible, Exposed, Infectious, Removed) model to surveillance of individuals by some governments as a result predict the spread. They also presented another interesting ap- of huge amount of the collected data [133]. Using the digital proach by pre-training an LSTM (Long Short-Term Memory) footprint data provided by the applications along with ML model which is a Recurrent Neural Network (RNN) model technology could allow users to identify infected patients and on data from SARS-CoV to predict the spread of COVID-19. enforce social distancing measures. Their findings discovered that both models achieved similar A real-time contact tracing using AI has been applied results in predicting the number of cases. in South Africa using the Sqreem platform [134], which Similarly, another ML model with clustering techniques is developed in Singapore in order to track people by the trained on data from the Chinese CDC, internet searches, and metadata of the device. This information is not including the news articles create a 2-day ahead real-time forecast about the personal data of the user. If the user enters an infected area, number of confirmed cases for 32 provinces in China [160]. he or she will be contacted by the authorities with respect While in some cases there are not enough COVID-19 data to the probability of infection. Mostly, contact tracing has available for accurate forecasting using ML techniques, Liu been conducted using a major accessible device, which is et al. [160] overcame the lack of data by implementing data a smartphone. A variety of smartphone applications enabled augmentation techniques. with ML or AI have been adopted in order to slow the spread Al et al. [161] introduced a novel forecasting technique of the virus by tracking and warning regarding the unsafe that allows them to predict the number of confirmed cases contacts [135]. Within the process of development, the very in China over the next 10 days. The authors combined and first part would be the consideration of framework, either modified a Flower Pollination Algorithm (FPA) [162] and a centralized or decentralized using appropriate technologies Salp Swam Algorithm (SSA) [163] to improve and evaluate such as Global Positioning System, Quick Response codes, the optimal parameters for an Adaptive Neuro-Fuzzy Inference and Bluetooth [136]. ML can enable alerting automatically System (ANFIS) [164] by creating an FPASSA-ANFIS model and analyzing the massive captured data, which would reduce that shows greater performance compared to other optimal the workforce [132]. Apple and Google announced that a parameters such as Root Mean Squared Relative Error (RMSE) Bluetooth-based platform for tracking close contacts will be or Mean Absolute Percentage Error (MAPE). implemented in the upcoming months. This technology will Alternatively, Rizk et al. [165] integrated algorithms and enable higher participation and better communication [137]. techniques like Interior Search Algorithm (ISA) and Chaotic An application is developed in South Korea in order to capture Learning (CL) into a Multi-Layer Feed-Forward Neural Net- the areas using location-based information, where a confirmed work (MFNN) creating a forecasting model called ISACL- case went before testing positive for COVID-19. Moreover, MFNN. Combining the two, CL and ISA, approaches im- text messages will be sent to people who may have been proved the overall performance of ISA. The authors retrieved exposed due to the contamination areas [138]. a dataset from WHO that included data from USA, Italy, Regarding the numerous implemented contact tracing ap- and Spain between January 2020 and April 2020. They plications, Table IV presents some of them that have been trained their model on this dataset and the model is then implemented in various countries [139]. assessed through the similarly aforementioned techniques such as RMSE and MAPE, and more. A new ML methodology GROOMs was proposed by Fong C. Forecasting et al. [166] for forecasting. The authors provided an ensemble Forecasting epidemics centers on tracking and predicting of forecasting and polynomial neural techniques that were the spread of infectious diseases and viruses. During an trained over the limited data to determine the technique that epidemic, forecasting methods and models can be trained on would yield with the lowest prediction error. Ayyoubzadeh epidemiological related data to provide an estimated number et al. [167] used Liner Regression (LR) and LSTM models of infected cases, patterns of spread that can give healthcare to predict the spread of COVID-19 in Iran. The authors workers guidance on how to prepare appropriately for an trained their model on data from Google Trends [168] and outbreak [157]. Previously statistical forecasting tool such as Worldometer [169]. They evaluated the model with 10-fold Susceptible, Infected, Recovered models (SIR Models) have cross-validation and use RMSE as a performance metric. been used to determine the spread of a disease through the Ghazaly et al. [170] utilized the limited data from WHO about population [158]. Recently, with the COVID-19 pandemic, confirmed COVID-19 cases and deaths between January 2020 TABLE IV CONTACT TRACING APPLICATIONS COMBATING COVID-19.

Reference Application Function Origin Technology Track close contacts of users Bluetooth [140] AarogyaSetu India Notifying user if captured users are infected GPS location Track close contacts of users GPS Display the situation of user by three colors Bank transactions’ history [141] Alipay Health Code China The situations include healthy, in need of short or long quarantine Tracking traveling information, and body temperature Track close contacts of users Bluetooth [142] BeAware Bahrain Bahrain Quarantined and self-isolated tracking Location Track close contacts of users [143] COVIDSafe Australia Bluetooth Notifying user if captured users are infected Track close contacts of users [144] CovTracer Cyprus GPS Location Notifying user if captured users are infected Track close contacts of users [145] CovidRadar Mexico Bluetooth Notifying user if captured users are infected Track close contacts of users Bluetooth [146] Ehteraz Qatar Notifying user if captured users are infected GPS Track close contacts of users [147] eRouska(CZ Smart Quarentine) Czech Republic Bluetooth Notifying user if captured users are infected Track the places an infected user had gone [148] GH Covid-19 Tracker App) Ghana GPS Allow for reporting symptoms [149] Hamagen Track close contacts of users Israel Location based on API Track close contacts of users [150] Immuni Italy Bluetooth Low Energy Notifying user if captured users are infected Measure the chance of infection [151] Ito Germany Bluetooth Guide for better safety manner Track close contacts of users [152] Mask.ir Provide a map of contaminated areas Iran Bluetooth Allow for reporting symptoms Track close contacts of users [153] MyTrace Malaysia Bluetooth Low Energy Notifying user if captured users are infected Track close contacts of users [154] StopCovid France Bluetooth Notifying user if captured users are infected Track close contacts of users [155] TraceCovid Accessing to the user’s information by government (privacy concern) UAE Bluetooth Notifying user if captured users are infected Track close contacts of users [156] TraceTogether Accessing to the user’s information by government (privacy concern) Singapore Bluetooth Notifying user if captured users are infected and April 2020 to train a Non-Linear Aggressive Model (NAR) Global Virus Tracker [173] built up a system embedded the and predict the future cases and deaths that could occur in 9 location and symptom risk evaluation for the users. They also countries. However, due to not having enough historical data developed the chatbot to provide the conversational interface for their model, the authors concluded their network is unable for the same ML-based evaluation. In the proposed system, to continue to predict the future. Over the same time period the county or city level risk, population density and updated of those three months [171] Roy et al. implemented their virus spreading were incorporated at the moment of the risk ML techniques to forecast the number of cases for infected, evaluation. recovered, and deceased cases country-wise and globally. The authors used a type of regression model called the Prophet D. Social Media Analysis Prediction Model. Developed by Facebook, this model has a Social media has become a platform where people share capacity to create a precise time-series forecast that is simple pictures, reviews, posts, and exchange stories. A popular social to create and could result in accurate prediction. media platform where people may obtain and access news is Bandyopadhyay et al. [172] used a RNN and GRU (Gated Twitter. Its users can validate live alerts and obtain information Recurrent Unit) model to predict the number of confirmed directly through the smartphone application. This is possible cases and deaths. The model was trained on data sourced as major news outlets, government bodies, community centers, from Kaggle on confirmed cases between January 2020 and etc., all have accounts that they use to share updates on March 2020. The results presented in their findings indicate Twitter. Users can also use the platform to share their personal that the RNN is capable of predicting the cases and assessing experiences via tweets. It can essentially be considered a the severity of COVID-19. Utilizing the ML techniques, the form of microblogging for users who just want to share their insight over a certain topic. Tweets can be a form of data for COVID-19. The immunologists around the world surged that can analyze feedback and obtain public sentiment over towards studying the symptoms, and the immune responses certain topics. Over the course of the COVID-19 pandemic, that infected patients show in relation to combating the virus people have engaged on twitter to share their experience with [184]. In this section, we review the efforts of the scientific regards to testing or lack of testing, social-distancing, and community in using ML techniques regarding understanding other challenges that people are facing due to the pandemic the virus [52], how to attack it, and perhaps even be able to [174]. In the remainder of the subsection, we review the find a cure for COVID-19. research dedicated on how ML technology can be used to analyze social media for COVID-19 related updates. A. Understanding the Virus The COVID-19 was declared a global pandemic in March Analyzing the genomics and proteomics characteristics of 2020. However, people had already been posting and dis- a viral disease is an important step to combat the disease. cussing it over social platforms. Between January 27th and Scientists have been studying the virology of COVID-19 March 26th, 2020 there were 5,621,048 tweets with keywords which can give the physical and chemical properties, cell “corona virus” and “coronavirus” according to Karisani et al. entry and receptor interaction, and the overall ecology and [175] when they constructed a dataset for their ML models. the genomics of the virus [185]. A genome is the complete The dataset consists of both positive and negative tweets genetic information that provides the architecture of a virus shared by users. Using this dataset, the authors were able to and knowing the genome for COVID-19 can help in better implement several ML methods like Logistic Regression (LR), understanding the transmissibility and infectiousness of the Naive Bayes Classification (NB) to automate the detection of virus [186]. The study of proteomics is knowing the proteins COVID-19 positive results shared over twitter by users. Both of an organism. Identifying the proteins of COVID-19 would LR and NB supervised ML models were also used by Samuel allow a better understanding of the overall protein structure et al. [176] to obtain public sentiment and feedback about and discover how the proteins would interact with the drugs COVID-19 through users tweets shared over the platform. As [187]. Over recent years, there have been remarkable advance- there is an abundance of news and tweets shared over twitter, ments by scientists in interdisciplinary fields of bioinformatics a sentiment analysis was done by [177] addressing the need and computational medicine and ML techniques have shown for filtering it out as there is a potential of misinformation meaningful interpretation towards determining genomics and being spread across social platforms. The authors implemented protein structures of various diseases [188]. In this section, we an unsupervised ML topic modelling technique known as focus on COVID-19 and discuss the ML techniques that have Latent Dirichlet Allocation (LDA). The use of LDA is done been implemented regarding the research of interpreting the to spot the semantic relationship between words in a tweet genomics and proteomics of that. and provides a sentiment analysis on whether the tweets COVID-19 is an RNA (ribonucleic acid) type of virus from are positive and showing signs of comfort or whether they the coronavirus (CoV) family. It is a single-stranded RNA with are negative showing discomfort and panic. In their findings, a large viral genome. These large genomes can have two or negative tweets are higher as people are prevalent in anger and three viral proteases. For COVID-19 it has 2 proteases, which sadness towards quarantine and death. we will refer to as 3CLpro. COVID-19 belongs to the same Jahanbin et al. [178] gathered data from Twitter by search- family as the aforementioned respiratory diseases SARS and ing for COVID-19 related hashtags. The dataset of tweets is MERS [189]. Viruses that belong to this type of family can pre-processed and filtered first to reduce the noise by removing infect a range of animal species such as camels, cats, cattle, irrelevant data. This would allow training a better model for bats, as well as humans [190]. They are easily transmittable classification. The authors used an evolutionary algorithm and can infect a host in one species and transmit it onto Eclass1-MIMO that predicts the morbidity rate in regions. another species [190]. Multiple findings suggest that the origin Mackey et al. [179] introduced an unsupervised ML approach of COVID-19 in infecting humans is transmitted from bats that analyzes tweets by users who may be infected by the with an 89% similarity structure identity to a coronavirus that virus, recovering from it or the experiences they had related to infects bats (SARS-like-CoVZXC21) [191]. testing for it. The authors used a Biterm Topic Model (BTM) Identifying and determining the biological structure at a combined with clustering techniques to determine statistical molecular level of a viral disease is important to develop and geographical characteristics depending on content analy- an effective therapy towards finding a cure for the disease. sis. Obtaining feedback over Twitter and social media outlets However, this process involves a lot of experiments that can will give a live reflection of how the general public is reacting take months. Computational ML techniques and models can to the pandemic and will allow policymakers in making better speed up this process and predict the structure of proteins decisions. accurately [192]. The family of coronavirus has multiple classes, and the IV. ML TECHNIQUES TOWARDS MEDICAL ASSISTANCE virus can belong to either the alpha or the beta class of As the virus spreads across the world infecting more of virus. SARS-CoV and MERS both were determined to be the population and with the death toll rising rapidly, efforts beta coronaviruses [193]. To determine and classify what are made to develop an effective vaccine or discover a drug specific type of virus COVID-19 could be, Randhawa, et al. TABLE V PREDICTIVE ANALYSIS TOOLSAND METHODS COMBATING COVID-19.

Reference Section Model and Technology Remarks [33] Predictive Analysis tool Using flight details data and recent outbreaks to predict the spread in nearby countries Early Tracking, Prevention [32] ANN - K-Means Algorithm Using a MAE to successfully predict 2-day spread [170] Non-Auto Regressive Neural Network Prediction Error, due to scarcity of error at the time of Analysis [160] Augmented ARGONet Clustering of Chinese Provinces, and getting a 2-day forecast [166] Polynomial Neural Network (PNN) - GROOMS Addressing data augmentation and importance of early forecast [172] RNN Researching predicting using GRU + LSTM combined models [180]Forecasting K-Means Clustering Algorithm Possible to predict the spread of cases [161] FPASSA-ANFIS (ANN) Predict a 10-day forecast of the number of cases in China [165] ISACL-MFNN Predict a 10-day forecast of the number of cases in multiple countries [167] LSTM and LR models Predict and forecast of the number of cases of COVID-19 in Iran [171] Regression Model, Prophet Prediction Time-Series Forecasting [181] AI Algorithms Phone based survey to determine whether a person is high-risk, low-risk or contracting the virus [178] Eclass1-MIMO Classifying a twitter dataset to determine morbidity in regions [176] Natural Processing Language Getting public sentiment by classifying tweets [177] Latent Dirichlet Allocation (LDA) Algorithms used to spot semantic relationship between words Social Media Analysis [182] Sentiment Analysis Building a visual cluster to highlighting public opinion over pandemic [179] Unsupervised ML (biterm topic model) Attain content analysis by assessing user tweets [175] Naive Bayes, Logistic Regression, and more. Automate detection of positive COVID-19 report results through tweets [183] Shallow Neural Networks Training multiple word2vec models to put context to words

[194] used a supervised ML technique with digital signal drugs or repurposing the currently approved FDA ones. We processing techniques (MLDSP) to evaluate the taxonomy of also review the ML research that has been done regarding the the virus. MLDSP techniques have previously been used to vaccine development. achieve high accuracy in the classification of other viruses and diseases such as HIV-1 genomes and Influenza [195]. Using 1) Drug Development and Repurposing: An exploratory this model, the authors could evaluate that like its predecessor, approach of determining whether commercially available anti- COVID-19 also belongs to the beta coronavirus. To predict viral drugs can treat or help towards reducing the severity the protein structure of COVID-19, Heo and Feig [196] used of COVID-19 infected patients was presented by Beck et a ML-based method called TrRosetta. This method can be al. [199]. They used a pre-trained ML learning model called used to predict the inter-residue distance and create structure Molecule Transformer-Drug Target Interaction (MT-DTI), an models for the protein. To do this with higher prediction interaction prediction model, to predict the binding affinity efficiency, the authors applied molecular dynamics simulation- between COVID-19 infected proteins and compounds. The based refinement. objective of their study was to identify potential FDA approved Magar et al. [197] highlighted the importance of biological drugs that may restrain the proteins of COVID-19. MT-DTI structure and protein sequences in combating the virus. The is capable of predicting the chemical sequences and amino authors developed an ML model to predict how synthetic anti- acid sequences of a target protein without the whole structure bodies inhibit the spread of COVID-19. The model can predict information. This is helpful to use as there was limited knowl- the response of these antibodies by understanding the binding edge on the overall structure of viral proteins of COVID-19 between the antibodies and the viral mutations. Training the initially. Regarding their advantage, the authors used the model on a dataset that includes virus anti-body sequences and MT-DTI model to predict binding affinities of 3,410 FDA- the clinical patient neutralization response allowed the authors approved drugs. Similarly, Heiser et al. [35] used proprietary to predict antibody responses. The authors used ML techniques DL techniques for the purpose of drug discovery. They used such as XGBoost, RF, Multilayer Perceptron (MLP), SVM, their model to evaluate how FDA and European Medicines and LR for their model. The XGBoost model achieves the Agency (EMA) approved drugs and compounds would affect highest accuracy in prediction over an 80%-20% split of train human cells, analyzing over 1,660 drugs. and test data. In some cases, various drugs from antiviral to antimalarial could be used to combat COVID-19 [200]. These drugs used B. Drug and Vaccine Development for combating severe illnesses are referred to as “parents” As COVID-19 cases continue to rise numbers of both the by Moskal et al. [200] in their study. The authors used ML infected cases and the death toll, it has become an urgent techniques like CNN, LSTM, and MLP analyzing the molec- need to discover a drug that could mitigate these numbers ular similarity between these “parents” drugs and second- from increasing any further. ML techniques can be used to generation drugs that could potentially be also used to fight analyze how drugs react to the viral proteins of COVID-19. against the virus. The authors introduced the second generation We have already seen ML methods and techniques like SVM drugs as “progeny”. This study is important in predicting other and Bayesian Classifiers being used for drug discovery and drugs that may be helpful in this pandemic. It can result in repurposing [198]. In this section we review the ML studies having a larger catalog of drugs, which provides alternative and research that had been done about discovering the new solutions if the “parent” drug fails to respond. To convert the molecular structure into a high-dimensional vector space, the that evaluate combinations of peptides and receptor cells. A Mol2Vec [201] method was used. program that OptiVax adopts is NetMHCPan [222], which uses Kadioglu et al. [202] identified three viral proteins as targets a feed-forward neural network. Thereby, peptide information for their ML approach. They targeted the Spike protein, the and receptor cell are input data that generate an affinity score nucleocapsid protein, and the 2’-o-ribose-methyltransferase to predict the binding as output. The prediction score is protein. The spike protein acts as a cellular receptor for the constantly improved as it uses the back-propagation method. host of the virus. The nucleocapsid protein plays a vital role EvalVax utilizes the genetic ancestry of three different cat- in coronavirus transcription and the overall forming of the egories of the population, i.e. white, black, and Asian. This genomics of the RNA virus. The 2’-o-ribose-methyltransferase data is used as an objective function to discover peptide and protein is an essential protein for coronavirus synthesis and receptor pairs for OptiVax. processing. The authors used ML algorithms against the three Similarly, Herst et al. [220] in their research about finding proteins in predicting how FDA-approved drugs and natural a vaccine employed a similar technique that was previously compounds react to the three proteins with such key charac- used for combating the Ebola virus. They used ML techniques teristics. like Artificial Neural Network (ANN), SVM, netMHC, and In [203] the authors used a DNN to predict and generate netMHCpan to predict potential vaccine candidates. a new small design for molecules that would be capable Ward et al. [215] mapped out the protein sequences of of inhibiting COVID-19 3CLpro. Targeting 3CLpro can be COVID-19. This data was used by the authors for prediction, an essential part with respect to the drugs development for specificity, and epitope analysis. Epitope is the molecule that COVID-19. Alternatively, Zhavoronkov et al. [204] utilized adds antibodies attached and is recognized by the immune 28 various types of ML methods such as Autoencoders, system. The authors used the ML-based program NetMHC- Generative Adversial Networks, Genetic Algorithms to predict Pan to locate the epitope sequences. With this, the authors and generate the molecular structures. Using deep-Q learning were able to create an online tool that would aid in epitope networks, Tang et al. [205] were able to generate 3CLpro prediction, coronavirus homology analysis, variation analysis, compounds of COVID-19. Being able to successfully predict and proteome analysis. Another study about epitope prediction these protein targets can provide advancements in developing was presented by Qiao et al. [219]. They employed DL a potential drug for the virus. Hu et al. [206] created an ML techniques that are able to predict the best epitope for peptide- model that predicts the binding between the potential drugs based COVID-19 vaccinations [219]. To do this, the authors and COVID-19 proteins. proposed a DL model that utilizes LSTM and RNN methods to capture sequence patterns of peptides and eventually be able to 2) Vaccine Development: Once a virus starts to spread and predict the new mutant antigens for patients. As an alternative turn into a global pandemic, there is a very little chance approach to predicting the epitope and protein sequence, an of stopping it without a vaccine [207]. That stands true for ML based tool called Ellipro was utilized by Rahman et al. COVID-19 as well. Historically, vaccination has been the [216]. The tool is capable of predicting and presenting a solution to control or slow the spread of a viral infection visual view of the protein sequence of the epitope within the [208] [209]. It is critical to have a vaccine developed to structure. Similarly Sarkar et al. [217] used the SVM method provide immunity against COVID-19 and stop this pandemic. to predict the toxic level of some epitopes. So far, the research for vaccine development of COVID-19 The study towards epitope prediction continues in the re- is dedicated with three different types of vaccines [210]. The search done by Prachar et al. [218] who employed various Whole Virus Vaccine represents a classical strategy for the techniques such as ANN and Position-Specific Weight Ma- development of vaccinations of viral disease. Subunit Vaccine trices (PSSM) algorithms to predict and verify COVID-19 relies on extracting the immune response against the S-spike epitopes. Ong et al. [213] introduced a vaccine designing protein for COVID-19 [210]. This will refrain it from docking approach referred to as Reverse Vaccinology (RV). The aim it with the hosts receptor protein [210]. The Nucleic Acid of RV is to identify a potential vaccine through bioinformatics Vaccine produces a protective immunological response to fight analysis of the pathogen genomes. They used Supervised ML against the virus by [211]. classifications such as LR, SVM, K-nearest Neighbor, RF, At present, there are over 22 vaccines that are at the clinical and Extreme Gradient Boosting (XGB) to train on annotated stage of trials in combating COVID-19 [212]. The process of proteins dataset with the objective of getting high prediction developing a vaccine would need to first go through the design accuracy for the protein candidates, the authors used an ML stage and then towards the testing and experimental stage on based tool referred as Vaxign-ML. animals and eventually in humans. In this section, we review the implicit research that is done over vaccine development V. FUTURE EXPECTATIONS and how ML techniques have been employed. Although we review many of the ML approaches regarding Ge et al. [214] used ML techniques to evaluate how small the impact of COVID-19 in this paper, there is still an virus strings (called peptides) bind to the human receptor essential need for developing solutions using ML to address cells. The authors created two ML programs OptiVax and the pandemic’s complications and challenges. Since there was EvalVax. Optivax combines different pre-existing programs no adopted method for fighting against COVID-19 when it TABLE VI VACCINEAND DRUG DEVELOPMENTOF COVID-19 USING MLALGORITHMS.

Reference Sections Model and Technology Remarks [213] ML Algorithms Logistic Regression, Support Vector Machine, etc Using a tool called Vaxign to implement Reverse Vaccinology [214] OptiVax, EvalVax, netMHCpan, etc. Predicting binding between virus proteins and Host cell Receptors [215] NetMHCPan Create an online tool for visualisation and extraction of COVID-19 meta-analysis [216]Vaccine Development Ellipro, multiple ML methods Predict the epitope structure [217] SVM Review epitope-based design for a COVID-19 vaccine [218] ANN Predicting COVID-19 epitopes [219] DeepNovo, LSTM, RNN Predictive analysis of protein sequences to discover antibodies in patients [220] netMHCpan, netMHC Use ML techniques to predict peptide sequences [35] DL Models - Neural Networks Analysing how approved FDA Drugs would work against COVID-19 [199] DL - Drug Target Interactions Re-purposing current drugs to find out whether there is an affinity between drug and proteins [202] Neural Networks and Naive Bayes Predicting drug interaction between proteins and compounds [200] CNN, LSTM and MLP models ML techniques to predict similarities between available drugs to combat COVID-19 Drug Re-purposing [206] Fine-Tuning AtomNet based Model Predicting binding between COVID-19 proteins and drug compounds [204] Various ML methods Use RL strategies to generate new 3CLpro structure [203] Deep Neural Network Creating small molecule interaction and targeting 3CLpro [221] Deep Learning Models Provide large scale virtual screening to identify protein interacting pairs was started, previous ML models regarding infectious diseases Assistance”. ML applications for each of these phases are (epidemiological models) can be helpful for the early stage primarily focused as such; “Screening” intended for diag- of COVID-19. As we discussed, detecting and screening nosing the virus through medical imaging data (COVID-19 COVID-19 using AI and ML techniques can play a key role related X-Rays and CT-Scans), “Tracking and Forecasting” in combating this pandemic. The combination of technologies towards forecasting and predicting the numbers of cases and like IoT devices with these techniques needs to be expanded contact tracing, and lastly, “Medical Assistance” with the aim for crowd areas including airports, subways or bus stations, of understanding the protein sequences and structure of the and so on. This development would enhance the identification virus and whether a cure could be found in combating it via a of suspicious cases within lesser both time and contamination. drug or vaccine. One of the main challenges that researchers It is important to implement an efficient method of achiev- face when diagnosing using ML techniques was the lack of ing high accuracy of detecting COVID-19 through medical relevant data that are made accessible to the public. Lack imaging and integrating ML techniques. It is essential to of data meant researchers had to use techniques like data overcome the challenges that are presented by this approach. augmentation, transfer learning, and fine-tuning models to Challenges such as lack of data and the privacy issue within improve prediction accuracy. Though these methods worked data collection, false information by media, the limited number well in some cases, more data would make these models more of expertise between AI and medical science. AI technologies robust. Similarly, forecasting models trained on more data for can also assist to implement the following expectations in predicting the spread and number of cases could be more the future: i) empowering the medical imaging devices using accurate. Regarding developing a vaccine or re-purposing, the noncontact automatic image capturing to prevent further it is important to have a good understanding of virology, infection from the patient to the radiologist or even another bioinformatics. Additionally, ML is especially important for patient, and ii) automatically monitoring the patients using researchers from different fields to collaborate and integrate intelligent video analysis. As the countries learn more about their knowledge in order to discover a cure. COVID-19, it is essential to have updated datasets. This can lead to better forecasting by implementing the proper ML REFERENCES models using those datasets. In addition to the forecasting concepts, investigating the effects of different social media [1] S. Latif, M. Usman, S. Manzoor, W. Iqbal, J. Qadir, G. Tyson, I. Castro, and their pathways for detecting early sign of possible future A. Razi, M. N. K. Boulos, A. Weller et al., “Leveraging data science pandemic. to combat COVID-19: A comprehensive review,” 2020. [2] L. Morawska and J. Cao, “Airborne transmission of SARS-CoV-2: The world should face the reality,” Environment International, p. 105730, VI.CONCLUSION 2020. Machine Learning (ML) models and techniques have vastly [3] CDC. Symptoms of coronavirus. [Online]. Avail- able: https://www.cdc.gov/coronavirus/2019-ncov/symptoms-testing/ been used in plenty of industries over the past decade. symptoms.html Within the healthcare industry, ML has been usually used [4] H. Nishiura, T. Kobayashi, T. Miyama, A. Suzuki, S.-m. Jung, for screening and diagnosing. In epidemiology area, ML is K. Hayashi, R. Kinoshita, Y. Yang, B. Yuan, A. R. Akhmetzhanov et al., basically utilized for forecasting and understanding epidemics “Estimation of the asymptomatic ratio of novel coronavirus infections (COVID-19),” International Journal of Infectious Diseases, vol. 94, p. and diseases. In this paper, we presented a comprehensive 154, 2020. survey of how ML applications have been used to fight against [5] J. A. Al-Tawfiq, “Asymptomatic coronavirus infection: MERS-CoV COVID-19. We presented the efforts that are taken by the ML and SARS-CoV-2 (COVID-19),” Travel medicine and infectious dis- ease, 2020. research communities to combat this virus across three main [6] G. News. Coronavirus (COVID-19). [Online]. Available: https: phases “Screening”, “Tracking and Forecasting” and “Medical //news.google.com/covid19/map?hl=en-US&gl=US&ceid=US:en [7] M. Nasajpour, S. Pouriyeh, R. M. Parizi, M. Dorodchi, M. Valero, and [30] S. Choi, J. Lee, M.-G. Kang, H. Min, Y.-S. Chang, and S. Yoon, H. R. Arabnia, “Internet of Things for current COVID-19 and future “Large-scale Machine Learning of media outlets for understanding pandemics: An exploratory study,” arXiv preprint arXiv:2007.11147, public reactions to nation-wide viral infection outbreaks,” Methods, 2020. vol. 129, pp. 50–59, 2017. [8] R. Vaishya, M. Javaid, I. H. Khan, and A. Haleem, “Artificial In- [31] S. Debnath, D. P. Barnaby, K. Coppa, A. Makhnevich, E. J. Kim, telligence (AI) applications for COVID-19 pandemic,” Diabetes & S. Chatterjee, V. Toth,´ T. J. Levy, M. d Paradis, S. L. Cohen et al., “Ma- Metabolic Syndrome: Clinical Research & Reviews, 2020. chine Learning to assist clinical decision-making during the COVID-19 [9] A. Alimadadi, S. Aryal, I. Manandhar, P. B. Munroe, B. Joe, and pandemic,” Bioelectronic Medicine, vol. 6, no. 1, pp. 1–8, 2020. X. Cheng, “Artificial Intelligence and Machine Learning to fight [32] Z. Hu, Q. Ge, L. Jin, and M. Xiong, “Artificial Intelligence forecasting COVID-19,” 2020. of COVID-19 in china,” arXiv preprint arXiv:2002.07112, 2020. [10] N. L. Bragazzi, H. Dai, G. Damiani, M. Behzadifar, M. Martini, and [33] Z. Allam, G. Dey, and D. S. Jones, “Artificial intelligence (AI) provided J. Wu, “How Big Data and Artificial Intelligence can help better man- early detection of the coronavirus (COVID-19) in China and will age the COVID-19 pandemic,” International Journal of Environmental influence future urban health policy internationally,” AI, vol. 1, no. 2, Research and Public Health, vol. 17, no. 9, p. 3176, 2020. pp. 156–165, 2020. [11] M. C. Chang and D. Park, “How can Blockchain help people in [34] U. Ozkaya, S. Ozturk, and M. Barstugan, “Coronavirus (COVID-19) the event of pandemics such as the COVID-19?” Journal of Medical classification using deep features fusion and ranking technique,” arXiv Systems, vol. 44, pp. 1–2, 2020. preprint arXiv:2004.03698, 2020. [12] M. E. Darnell, K. Subbarao, S. M. Feinstone, and D. R. Taylor, [35] K. Heiser, P. F. McLean, C. T. Davis, B. Fogelson, H. B. Gordon, “Inactivation of the coronavirus that induces severe acute respiratory P. Jacobson, B. L. Hurst, B. J. Miller, R. W. Alfa, B. A. Earnshaw et al., syndrome, SARS-CoV,” Journal of Virological Methods, vol. 121, “Identification of potential treatments for COVID-19 through artificial no. 1, pp. 85–91, 2004. intelligence-enabled phenomic analysis of human cells infected with [13] M. Willman, D. Kobasa, and J. Kindrachuk, “A comparative analysis SARS-CoV-2,” bioRxiv, 2020. of factors influencing two outbreaks of middle eastern respiratory [36] V. M. Corman, O. Landt, M. Kaiser, R. Molenkamp, A. Meijer, D. K. syndrome (MERS) in Saudi Arabia and South Korea,” Viruses, vol. 11, Chu, T. Bleicker, S. Brunink,¨ J. Schneider, M. L. Schmidt et al., no. 12, p. 1119, 2019. “Detection of 2019 novel coronavirus (2019-nCoV) by real-time rt- [14] WHO. SARS - WHO. [Online]. Available: https://www.who.int/ith/ pcr,” Eurosurveillance, vol. 25, no. 3, p. 2000045, 2020. diseases/sars/en/ [37] X. Xie, Z. Zhong, W. Zhao, C. Zheng, F. Wang, and J. Liu, “Chest [15] L. Jia, Q. Zhao, and L. Tong, “Retail pricing for stochastic demand ct for typical 2019-nCoV pneumonia: relationship to negative rt-pcr with unknown parameters: An online Machine Learning approach,” in testing,” Radiology, p. 200343, 2020. 2013 51st Annual Allerton Conference on Communication, Control, [38] C. Gollier and O. Gossner, “Group testing against COVID-19,” Covid and Computing (Allerton). IEEE, 2013, pp. 1353–1358. Economics, vol. 2, 2020. [16] K. Chitra and B. Subashini, “Data Mining techniques and its applica- [39] Q.-V. Pham, D. C. Nguyen, W.-J. Hwang, P. N. Pathirana et al., tions in banking sector,” International Journal of Emerging Technology “Artificial Intelligence (AI) and Big Data for coronavirus (COVID-19) and Advanced Engineering, vol. 3, no. 8, pp. 219–226, 2013. pandemic: A survey on the state-of-the-arts,” 2020. [17] M. Chen, Y. Hao, K. Hwang, L. Wang, and L. Wang, “Disease predic- tion by Machine Learning over big data from healthcare communities,” [40] S. Lalmuanawma, J. Hussain, and L. Chhakchhuak, “Applications of IEEE Access, vol. 5, pp. 8869–8879, 2017. Machine Learning and Artificial Intelligence for COVID-19 (SARS- CoV-2) pandemic: A review,” Chaos, Solitons & Fractals, p. 110059, [18] J. Clemente, F. Li, M. Valero, and W. Song, “Smart seismic sensing 2020. for indoor fall detection, location, and notification,” IEEE journal of biomedical and health informatics, vol. 24, no. 2, pp. 524–532, 2020. [41] W. Yang and F. Yan, “Patients with rt-pcr-confirmed COVID-19 and Radiology [19] S. Ekins, “The next era: Deep Learning in pharmaceutical research,” normal chest ct,” , vol. 295, no. 2, pp. E3–E3, 2020. Pharmaceutical Research, vol. 33, no. 11, pp. 2594–2603, 2016. [42] S. H. Kassani, P. H. Kassasni, M. J. Wesolowski, K. A. Schneider, and [20] I. El Naqa and M. J. Murphy, “What is Machine Learning?” pp. 3–11, R. Deters, “Automatic detection of coronavirus disease (COVID-19) 2015. in x-ray and ct images: A Machine Learning-based approach,” arXiv preprint arXiv:2004.10641, 2020. [21] E. Alpaydin, Introduction to Machine Learning. MIT press, 2020. [22] S. Wong, H. Al-Hasani, Z. Alam, and A. Alam, “Artificial Intelligence [43] A. Giovagnoni, “Facing the COVID-19 emergency: we can and we in radiology: how will we be affected?” European Radiology, vol. 29, do.” La Radiologia Medica, pp. 1–2, 2020. no. 1, pp. 141–143, 2019. [44] D. Cozzi, M. Albanesi, E. Cavigli, C. Moroni, A. Bindi, S. Luvara,` [23] A. D. T. Force*., M. Singer, C. Deutschman, C. Seymour, M. Wildman, S. Lucarini, S. Busoni, L. N. Mazzoni, and V. Miele, “Chest x-ray C. Sanderson, J. Groves, M. Wildman, C. Sanderson, J. Groves et al., in new coronavirus disease 2019 (COVID-19) infection: findings and “Artificial Intelligence in healthcare: past, present and future,” Journal correlation with clinical outcome,” La Radiologia Medica, p. 1, 2020. of the Intensive Care Society, vol. 20, no. 3, pp. 268–273, 2019. [45] A. E. Hassanien, L. N. Mahdy, K. A. Ezzat, H. H. Elmousalami, [24] K. Kourou, T. P. Exarchos, K. P. Exarchos, M. V. Karamouzis, and and H. A. Ella, “Automatic x-ray COVID-19 lung image classification D. I. Fotiadis, “Machine Learning applications in cancer prognosis system based on multi-level thresholding and support vector machine,” and prediction,” Computational and Structural Biotechnology Journal, medRxiv, 2020. vol. 13, pp. 8–17, 2015. [46] P. K. Sethy and S. K. Behera, “Detection of coronavirus disease [25] Q. Zou, K. Qu, Y. Luo, D. Yin, Y. Ju, and H. Tang, “Predicting diabetes (COVID-19) based on deep features,” Preprints, vol. 2020030300, p. mellitus with Machine Learning techniques,” Frontiers in Genetics, 2020, 2020. vol. 9, p. 515, 2018. [47] I. Castiglioni, D. Ippolito, M. Interlenghi, C. B. Monti, C. Salvatore, [26] C.-C. Wu, W.-C. Yeh, W.-D. Hsu, M. M. Islam, P. A. A. Nguyen, T. N. S. Schiaffino, A. Polidori, D. Gandola, C. Messa, and F. Sardanelli, Poly, Y.-C. Wang, H.-C. Yang, and Y.-C. J. Li, “Prediction of fatty liver “Artificial Intelligence applied on chest x-ray can aid in the diagnosis of disease using Machine Learning algorithms,” Computer Methods and COVID-19 infection: a first experience from lombardy, Italy,” medRxiv, Programs in Biomedicine, vol. 170, pp. 23–29, 2019. 2020. [27] H. Asri, H. Mousannif, H. Al Moatassime, and T. Noel, “Using [48] J. Zhang, Y. Xie, Y. Li, C. Shen, and Y. Xia, “COVID-19 screening Machine Learning algorithms for breast cancer risk prediction and on chest x-ray images using deep learning based anomaly detection,” diagnosis,” Procedia Computer Science, vol. 83, pp. 1064–1069, 2016. arXiv preprint arXiv:2003.12338, 2020. [28] M. A. Ahmad, C. Eckert, and A. Teredesai, “Interpretable Machine [49] L. Wang and A. Wong, “COVID-Net: A tailored deep convolutional Learning in healthcare,” in Proceedings of the 2018 ACM international neural network design for detection of COVID-19 cases from chest conference on bioinformatics, computational biology, and health infor- x-ray images,” arXiv preprint arXiv:2003.09871, 2020. matics, 2018, pp. 559–560. [50] I. D. Apostolopoulos and T. A. Mpesiana, “COVID-19: automatic de- [29] R. Sandhu, S. K. Sood, and G. Kaur, “An intelligent system for tection from x-ray images utilizing transfer learning with Convolutional predicting and preventing MERS-CoV infection outbreak,” The Journal Neural Networks,” Physical and Engineering Sciences in Medicine, of Supercomputing, vol. 72, no. 8, pp. 3033–3056, 2016. p. 1, 2020. [51] B. Ghoshal and A. Tucker, “Estimating uncertainty and interpretability [73] S. Minaee, R. Kafieh, M. Sonka, S. Yazdani, and G. J. Soufi, “Deep- in deep learning for coronavirus (COVID-19) detection,” arXiv preprint COVID: Predicting COVID-19 from chest x-ray images using deep arXiv:2003.10769, 2020. transfer learning,” arXiv preprint arXiv:2004.09363, 2020. [52] J. Chen, K. Li, Z. Zhang, K. Li, and P. S. Yu, “A survey on applications [74] D. Das, K. Santosh, and U. Pal, “Truncated inception net: COVID- of Artificial Intelligence in fighting against COVID-19,” arXiv preprint 19 outbreak screening using chest x-rays,” Physical and engineering arXiv:2007.02202, 2020. sciences in medicine, pp. 1–11, 2020. [53] M. Farooq and A. Hafeez, “COVID-resnet: A Deep Learning frame- [75] S. Basu and S. Mitra, “Deep Learning for screening COVID-19 using work for screening of COVID19 from radiographs,” arXiv preprint chest x-ray images,” arXiv preprint arXiv:2004.10507, 2020. arXiv:2003.14395, 2020. [76] M. Rahimzadeh and A. Attar, “A new modified Deep Convolutional [54] E. E.-D. Hemdan, M. A. Shouman, and M. E. Karar, “COVIDx-net: Neural Network for detecting COVID-19 from x-ray images,” arXiv A framework of Deep Learning classifiers to diagnose COVID-19 in preprint arXiv:2004.08052, 2020. x-ray images,” arXiv preprint arXiv:2003.11055, 2020. [77] Y. Zhang, S. Niu, Z. Qiu, Y. Wei, P. Zhao, J. Yao, J. Huang, Q. Wu, and [55] H. Hirano, K. Koga, and K. Takemoto, “Vulnerability of deep neural M. Tan, “COVID-da: Deep domain adaptation from typical pneumonia networks for detecting COVID-19 cases from chest x-ray images to to COVID-19,” arXiv preprint arXiv:2005.01577, 2020. universal adversarial attacks,” arXiv preprint arXiv:2005.11061, 2020. [78] I. D. Apostolopoulos, S. I. Aznaouridis, and M. A. Tzani, “Extract- [56] P. Afshar, S. Heidarian, F. Naderkhani, A. Oikonomou, K. N. Platan- ing possibly representative COVID-19 biomarkers from x-ray images iotis, and A. Mohammadi, “COVID-caps: A capsule network-based with Deep Learning approach and image data related to pulmonary framework for identification of COVID-19 cases from x-ray images,” diseases,” Journal of Medical and Biological Engineering, p. 1, 2020. arXiv preprint arXiv:2004.02696, 2020. [79] H. Choi, X. Qi, S. H. Yoon, S. J. Park, K. H. Lee, J. Y. Kim, Y. K. Lee, [57] L. Sarker, M. M. Islam, T. Hannan, and Z. Ahmed, “COVID-DenseNet: H. Ko, K. H. Kim, C. M. Park et al., “Extension of coronavirus disease A Deep Learning architecture to detect COVID-19 from chest radiology 2019 (COVID-19) on chest ct and implications for chest radiograph images,” 2020. interpretation,” Radiology: Cardiothoracic Imaging, vol. 2, no. 2, p. [58] F. Ucar and D. Korkmaz, “COVIDiagnosis-Net: Deep Bayes- e200107, 2020. SqueezeNet based diagnostic of the coronavirus disease 2019 (COVID- [80] H. Shi, X. Han, N. Jiang, Y. Cao, O. Alwalid, J. Gu, Y. Fan, and 19) from x-ray images,” Medical Hypotheses, p. 109761, 2020. C. Zheng, “Radiological findings from 81 patients with COVID- [59] S. Khobahi, C. Agarwal, and M. Soltanalian, “CoroNet: A deep 19 pneumonia in Wuhan, China: a descriptive study,” The Lancet network architecture for semi-supervised task-based identification of Infectious Diseases, 2020. COVID-19 from chest x-ray images,” medRxiv, 2020. [81] A. A. Ardakani, A. R. Kanafi, U. R. Acharya, N. Khadem, and [60] E. Luz, P. L. Silva, R. Silva, and G. Moreira, “Towards an efficient A. Mohammadi, “Application of Deep Learning technique to manage Deep Learning model for COVID-19 patterns detection in x-ray im- COVID-19 in routine clinical practice using ct images: Results of 10 ages,” arXiv preprint arXiv:2004.05717, 2020. convolutional neural networks,” Computers in Biology and Medicine, [61] A. Narin, C. Kaya, and Z. Pamuk, “Automatic detection of coronavirus p. 103795, 2020. disease (COVID-19) using x-ray images and Deep Convolutional [82] L. Li, L. Qin, Z. Xu, Y. Yin, X. Wang, B. Kong, J. Bai, Y. Lu, Z. Fang, Neural Networks,” arXiv preprint arXiv:2003.10849, 2020. Q. Song et al., “Artificial Intelligence distinguishes COVID-19 from [62] H. S. Maghdid, A. T. Asaad, K. Z. Ghafoor, A. S. Sadiq, and M. K. community acquired pneumonia on chest ct,” Radiology, p. 200905, Khan, “Diagnosing COVID-19 pneumonia from x-ray and ct images 2020. using Deep Learning and transfer learning algorithms,” arXiv preprint [83] F. Shan, Y. Gao, J. Wang, W. Shi, N. Shi, M. Han, Z. Xue, and Y. Shi, arXiv:2004.00038, 2020. “Lung infection quantification of COVID-19 in ct images with Deep [63] M. Togac¸ar,˘ B. Ergen, and Z. Comert,¨ “COVID-19 detection using Learning,” arXiv preprint arXiv:2003.04655, 2020. Deep Learning models to exploit social mimic optimization and struc- [84] A. Holzinger, M. Plass, M. Kickmeier-Rust, K. Holzinger, G. C. tured chest x-ray images using fuzzy color and stacking approaches,” Cris¸an, C.-M. Pintea, and V. Palade, “Interactive machine learning: Computers in Biology and Medicine, p. 103805, 2020. experimental evidence for the human in the algorithmic loop,” Applied [64] T. Ozturk, M. Talo, E. A. Yildirim, U. B. Baloglu, O. Yildirim, Intelligence, vol. 49, no. 7, pp. 2401–2414, 2019. and U. R. Acharya, “Automated detection of COVID-19 cases using [85] Z. Tang, W. Zhao, X. Xie, Z. Zhong, F. Shi, J. Liu, and D. Shen, “Sever- deep neural networks with x-ray images,” Computers in Biology and ity assessment of coronavirus disease 2019 (COVID-19) using quanti- Medicine, p. 103792, 2020. tative features from chest ct images,” arXiv preprint arXiv:2003.11988, [65] S. Rajaraman and S. Antani, “Training Deep Learning algorithms with 2020. weakly labeled pneumonia chest x-ray data for COVID-19 detection,” [86] O. Gozes, M. Frid-Adar, H. Greenspan, P. D. Browning, H. Zhang, medRxiv, 2020. W. Ji, A. Bernheim, and E. Siegel, “Rapid ai development cycle for [66] D. N. Vinod and S. Prabaharan, “Data science and the role of Artificial the coronavirus (COVID-19) pandemic: Initial results for automated Intelligence in achieving the fast diagnosis of COVID-19,” Chaos, detection & patient monitoring using Deep Learning ct image analysis,” Solitons & Fractals, p. 110182, 2020. arXiv preprint arXiv:2003.05037, 2020. [67] M. Jamil, I. Hussain et al., “Automatic detection of COVID-19 infec- [87] J. Zhao, Y. Zhang, X. He, and P. Xie, “COVID-ct-dataset: a ct scan tion from chest x-ray using Deep Learning,” medRxiv, 2020. dataset about COVID-19,” arXiv preprint arXiv:2003.13865, 2020. [68] S. Toraman, T. B. Alakus¸, and I.˙ Turko¨ glu,˘ “Convolutional capsnet: [88] C. Butt, J. Gill, D. Chun, and B. A. Babu, “Deep Learning system A novel artificial neural network approach to detect COVID-19 dis- to screen coronavirus disease 2019 pneumonia,” Applied Intelligence, ease from x-ray images using capsule networks,” Chaos, Solitons & p. 1, 2020. Fractals, p. 110122, 2020. [89] S. Wang, B. Kang, J. Ma, X. Zeng, M. Xiao, J. Guo, M. Cai, J. Yang, [69] M. E. Chowdhury, T. Rahman, A. Khandakar, R. Mazhar, M. A. Kadir, Y. Li, X. Meng et al., “A Deep Learning algorithm using ct images to Z. B. Mahbub, K. R. Islam, M. S. Khan, A. Iqbal, N. Al-Emadi et al., screen for corona virus disease (COVID-19),” MedRxiv, 2020. “Can AI help in screening viral and COVID-19 pneumonia?” arXiv [90] X. Bai, C. Fang, Y. Zhou, S. Bai, Z. Liu, L. Xia, Q. Chen, Y. Xu, preprint arXiv:2003.13145, 2020. T. Xia, S. Gong et al., “Predicting COVID-19 malignant progression [70] H. Panwar, P. Gupta, M. K. Siddiqui, R. Morales-Menendez, and with AI techniques,” 2020. V. Singh, “Application of Deep Learning for fast detection of COVID- [91] Y. Pathak, P. K. Shukla, A. Tiwari, S. Stalin, S. Singh, and P. K. Shukla, 19 in x-rays using nCOVnet,” Chaos, Solitons & Fractals, p. 109944, “Deep Transfer Learning based classification model for COVID-19 2020. disease,” IRBM, 2020. [71] B. Hurt, S. Kligerman, and A. Hsiao, “Deep Learning localization [92] S. Jin, B. Wang, H. Xu, C. Luo, L. Wei, W. Zhao, X. Hou, W. Ma, of pneumonia: 2019 coronavirus (COVID-19) outbreak,” Journal of Z. Xu, Z. Zheng et al., “Ai-assisted ct imaging analysis for COVID-19 Thoracic Imaging, vol. 35, no. 3, pp. W87–W89, 2020. screening: Building and deploying a medical AI system in four weeks,” [72] D. Ezzat, H. A. Ella et al., “GSA-DenseNet121-COVID-19: a hybrid medRxiv, 2020. deep learning architecture for the diagnosis of COVID-19 disease [93] C. Jin, W. Chen, Y. Cao, Z. Xu, X. Zhang, L. Deng, C. Zheng, J. Zhou, based on gravitational search optimization algorithm,” arXiv preprint H. Shi, and J. Feng, “Development and evaluation of an AI system for arXiv:2004.05084, 2020. COVID-19 diagnosis,” medRxiv, 2020. [94] G. Wang, X. Liu, C. Li, Z. Xu, J. Ruan, H. Zhu, T. Meng, K. Li, data-available-to-aid-covid-19-efforts/, 2020, Accessed August 04, N. Huang, and S. Zhang, “A noise-robust framework for automatic 2020. segmentation of COVID-19 pneumonia lesions from ct images,” IEEE [114] https://www.cdc.gov/coronavirus/2019-ncov/symptoms-testing/ Transactions on Medical Imaging, 2020. symptoms.html, 2020, Accessed August 04, 2020. [95] J. Song, H. Wang, Y. Liu, W. Wu, G. Dai, Z. Wu, P. Zhu, W. Zhang, [115] https://www.symptoma.com/, 2020, Accessed August 04, 2020. K. W. Yeom, and K. Deng, “End-to-end automatic differentiation of [116] Symptoma, https://www.symptoma.com/en/about, 2020, Accessed Au- the coronavirus disease 2019 (COVID-19) from viral pneumonia based gust 06, 2020. on chest ct,” European journal of nuclear medicine and molecular [117] H. HaddadPajouh, A. Dehghantanha, R. M. Parizi, M. Aledhari, and imaging, pp. 1–9, 2020. H. Karimipour, “A survey on Internet of Things security: Requirements, [96] Y.-H. Wu, S.-H. Gao, J. Mei, J. Xu, D.-P. Fan, C.-W. Zhao, and M.-M. challenges, and solutions,” Internet of Things, p. 100129, 2019. Cheng, “JCS: An explainable COVID-19 diagnosis system by joint [118] M. Sheng, Y. Qin, L. Yao, and B. Benatallah, Eds., Managing the Web classification and segmentation,” arXiv preprint arXiv:2004.07054, of Things: Linking the Real World to the Web. Morgan Kaufmann, 2020. 2017. [97] T. Zhou, S. Canu, and S. Ruan, “An automatic COVID-19 ct seg- [119] A. H. Ngu, M. Gutierrez, V. Metsis, S. Nepal, and Q. Z. Sheng, “IoT mentation based on U-Net with attention mechanism,” arXiv preprint middleware: A survey on issues and enabling technologies,” IEEE arXiv:2004.06673, 2020. Internet of Things Journal, vol. 4, no. 1, pp. 1–20, 2017. [98] K. Yang, X. Liu, Y. Yang, X. Liao, R. Wang, X. Zeng, Y. Wang, [120] D. Spiliopoulos, “How iot and ai can enable companies to restart their M. Zhang, and T. Zhang, “End-to-end COVID-19 screening with 3D business during COVID-19,” https://iottechnews.com/news/2020/jun/ Deep Learning on chest computed tomography,” 2020. 02/how-iot-and-ai-can-enable-companies-during-covid-19-to-restart- [99] S. Yang, L. Jiang, Z. Cao, L. Wang, J. Cao, R. Feng, Z. Zhang, X. Xue, their-business/, 2020, Accessed August 07, 2020. Y. Shi, and F. Shan, “Deep learning for detecting corona virus disease [121] SmartXHub, “Automatic fever screening tablet,” https: 2019 (COVID-19) on high-resolution computed tomography: a pilot //smartxhub.com/tablet-temperature-screening-brochure/, 2020, study,” Annals of Translational Medicine, vol. 8, no. 7, 2020. Accessed August 07, 2020. [100] A. Mobiny, P. A. Cicalese, S. Zare, P. Yuan, M. Abavisani, C. C. [122] G. Onag, “Taiwan hospital taps AI, IoT and cloud to keep Wu, J. Ahuja, P. M. de Groot, and H. Van Nguyen, “Radiologist- away COVID-19,” https://futureiot.tech/taiwan-hospital-taps-ai-iot- level COVID-19 detection using ct scans with detail-oriented capsule and-cloud-to-keep-away-covid-19/, 2020, Accessed August 08, 2020. networks,” arXiv preprint arXiv:2004.07407, 2020. [123] R. Sagar, “How Computer Vision came in handy for social distanc- [101] Y. Song, S. Zheng, L. Li, X. Zhang, X. Zhang, Z. Huang, J. Chen, ing,” https://analyticsindiamag.com/covid-19-computer-vision/, 2020, H. Zhao, Y. Jie, R. Wang et al., “Deep learning enables accurate Accessed August 08, 2020. diagnosis of novel coronavirus (COVID-19) with ct images,” medRxiv, [124] “AI powered workplace safety and compliance,” http://rayvision.ai, 2020. 2020, Accessed August 08, 2020. [102] C. Zheng, X. Deng, Q. Fu, Q. Zhou, J. Feng, H. Ma, W. Liu, and [125] K. Waddell, “Automating humans with AI,” https://www.axios.com/ X. Wang, “Deep learning-based detection for COVID-19 from chest ct ai-employee-surveillance-automating-humans-092f8c36-24bc-4efd- using weak label,” medRxiv, 2020. 8d38-78f1c8a61e1f.html, 2019, Accessed August 08, 2020. [103] B. A. Shawar and E. Atwell, “Chatbots: are they really useful?” in Ldv [126] “Landing AI creates an AI tool to help customers monitor social dis- forum, vol. 22, no. 1, 2007, pp. 29–49. tancing in the workplace,” https://landing.ai/landing-ai-creates-an-ai- [104] A. Martin, J. Nateqi, S. Gruarin, N. Munsch, I. Abdarahmane, and tool-to-help-customers-monitor-social-distancing-in-the-workplace/, B. Knapp, “An Artificial Intelligence-based first-line defence against 2020, Accessed August 08, 2020. COVID-19: digitally screening citizens for risks via a chatbot,” bioRxiv, [127] L. Ramadass, S. Arunachalam, and Z. Sagayasree, “Applying Deep 2020. Learning algorithm to maintain social distance in public place through [105] J. Cahn, “Chatbot: Architecture, design, & development,” University of drone technology,” International Journal of Pervasive Computing and Pennsylvania School of Engineering and Applied Science Department Communications, 2020. of Computer and Information Science, 2017. [128] A. Rai, “Explainable AI: from black box to glass box,” Journal of the [106] T. J. Judson, A. Y. Odisho, J. J. Young, O. Bigazzi, D. Steuer, Academy of Marketing Science, vol. 48, no. 1, pp. 137–141, 2020. R. Gonzales, and A. B. Neinstein, “Case report: Implementation of a [129] K. Ganasegeran and S. A. Abdulrahman, “Artificial Intelligence ap- digital chatbot to screen health system employees during the COVID-19 plications in tracking health behaviors during disease epidemics,” in pandemic,” Journal of the American Medical Informatics Association, Human Behaviour Analysis Using Intelligent Systems. Springer, 2020, 2020. pp. 141–155. [107] A. S. Miner, L. Laranjo, and A. B. Kocaballi, “Chatbots in the fight [130] W. Naude,´ “Artificial Intelligence against COVID-19: An early review,” against the COVID-19 pandemic,” npj Digital Medicine, vol. 3, no. 1, 2020. pp. 1–4, 2020. [131] CDC, “What you should know about COVID-19 to protect yourself and [108] U. Bharti, D. Bajaj, H. Batra, S. Lalit, S. Lalit, and A. Gangwani, others,” https://www.cdc.gov/coronavirus/2019-ncov/downloads/2019- “Medbot: Conversational Artificial Intelligence powered chatbot for ncov-factsheet.pdf, 2020, Accessed August 08, 2020. delivering tele-health after COVID-19,” in 2020 5th International Con- [132] T. Coles, “Contact tracing apps use ML to curb COVID- ference on Communication and Electronics Systems (ICCES). IEEE, 19 outbreaks,” https://www.itprotoday.com/machine-learning/contact- 2020, pp. 870–875. tracing-apps-use-ml-curb-covid-19-outbreaks, 2020, Accessed August [109] N. Ouerhani, A. Maalel, H. B. Ghezala,´ and S. Chouri, “Smart ubiqui- 08, 2020. tous chatbot for COVID-19 assistance with Deep Learning sentiment [133] R. A. Calvo, S. Deterding, and R. M. Ryan, “Health surveillance during analysis model during and after quarantine,” 2020. COVID-19 pandemic,” 2020. [110] Bespoke, “Bebot launches free coronavirus information bot,” https:// [134] A. N. Desk, “Tracing a million steps: SQREEM launches AI-driven www.be-spoke.io/index.html, 2020, Accessed August 06, 2020. contact tracing and communications platform to fight COVID-19,” [111] Orbita, “Orbita launches COVID-19 chatbot screener & https://aithority.com/technology/analytics/tracing-a-million-steps- knowledge base,” https://go.orbita.ai/orbita-launches-screening- sqreem-launches-ai-driven-contact-tracing-and-communications- chatbot-covid19-patient-education-triage-virtual-navigation?utm platform-to-fight-covid-19/, 2020, Accessed August 10, 2020. campaign=[Employee%20Health%20Manager%20-%20Orbita% [135] L. J. Kricka, S. Polevikov, J. Y. Park, P. Fortina, S. Bernardini, 20COVID-19]%20April%202020%20-%20present&utm source= D. Satchkov, V. Kolesov, and M. Grishkov, “Artificial Intelligence- orbitahomepagebody&utm content=covid19landingpage, 2020, powered search tools and resources in the fight against COVID-19,” Accessed August 06, 2020. EJIFCC, vol. 31, no. 2, p. 106, 2020. [112] Hyro, “A free virtual assistant to support health enterprises and their [136] J. Li and X. Guo, “COVID-19 contact-tracing apps: A survey on the patients,” https://www.hyro.ai/covid-19, 2020, Accessed August 06, global deployment and challenges,” arXiv preprint arXiv:2005.03599, 2020. 2020. [113] “Apple makes mobility data available to aid COVID-19 efforts,” [137] M. Panzzarino, “Apple and google are launching a joint COVID-19 https://www.apple.com/newsroom/2020/04/apple-makes-mobility- tracing tool for iOS and Android,” Tech Crunch, vol. 10, 2020. [138] S. Kwon, “COVID-19: Lessons from South Korea,” https: [165] R. M. Rizk-Allah and A. E. Hassanien, “COVID-19 forecasting based //healthsystemsglobal.org/news/covid-19-lessons-from-south-korea/, on an improved interior search algorithm and multi-layer feed forward 2020, Accessed August 18, 2020. neural network,” arXiv preprint arXiv:2004.05960, 2020. [139] P. Howell O’Neill, T. Ryan-Mosley, and B. Johnson, “A flood [166] S. J. Fong, G. Li, N. Dey, R. G. Crespo, and E. Herrera-Viedma, of coronavirus apps are tracking us. now its time to keep track “Finding an accurate early forecasting model from small dataset: of them.” https://www.technologyreview.com/2020/05/07/1000961/ A case of 2019-ncov novel coronavirus outbreak,” arXiv preprint launching-mittr-covid-tracing-tracker/, 2020, Accessed August 18, arXiv:2003.10776, 2020. 2020. [167] S. M. Ayyoubzadeh, S. M. Ayyoubzadeh, H. Zahedi, M. Ahmadi, and [140] “Aarogya Setu mobile app,” https://www.mygov.in/aarogya-setu-app/, S. R. N. Kalhori, “Predicting COVID-19 incidence through analysis 2020, Accessed August 16, 2020. of google trends data in Iran: Data Mining and Deep Learning pilot [141] “Alipay app,” https://mp.weixin.qq.com/s/ study,” JMIR Public Health and Surveillance, vol. 6, no. 2, p. e18828, amB7fBxLw8KSR9DcUsbTWg, 2020, Accessed August 17, 2020. 2020. [142] “BeAware Bahrain,” https://apps.bahrain.bh/CMSWebApplication/ [168] Google. Coronavirus search trends. [Online]. Available: https: action/ShowAppDetailsAction?selectedAppID=321&appLanguage=en, //trends.google.com/trends/story/US cu 4Rjdh3ABAABMHM en 2020, Accessed August 16, 2020. [169] WorldPop. Worldpop COVID-19. [Online]. Available: https: [143] “COVIDSafe app,” https://www.health.gov.au/resources/apps-and- //www.worldpop.org/covid19 tools/covidsafe-app#about-the-app, 2020, Accessed August 16, 2020. [170] N. M. Ghazaly, M. A. Abdel-Fattah, and A. Abd El-Aziz, “Novel coronavirus forecasting model using nonlinear autoregressive artificial [144] “COVTRACER,” https://covid-19.rise.org.cy/en/, 2020, Accessed Au- neural network,” Journal of Advanced Science, 2020. gust 16, 2020. [171] A. N. Roy, J. Jose, A. Sunil, N. Gautam, D. Nathalia, and A. Suresh, [145] “CovidRadar.mx,” https://covidradar.mx, 2020, Accessed August 16, “Prediction and spread visualization of COVID-19 pandemic using 2020. machine learning,” 2020. [146] “Ehteraz app made compulsory for citizens, residents,” [172] S. K. Bandyopadhyay and S. Dutta, “Machine Learning approach for https://www.marhaba.qa/qatars-covid-19-track-and-control-app- confirmation of COVID-19 cases: Positive, negative, death and release,” available-for-download-in-app-store/, 2020, Accessed August 16, medRxiv, 2020. 2020. [173] M. Han, “Global virus tracker,” https://globalvirustracker.org/, 2020, [147] “I protect both you and me,” https://erouska.cz/en, 2020, Accessed Accessed August 20, 2020. August 16, 2020. [174] M. Cinelli, W. Quattrociocchi, A. Galeazzi, C. M. Valensise, E. Brug- [148] “Gh COVID-19 tracker,” https://ghcovid19.com, 2020, Accessed Au- noli, A. L. Schmidt, P. Zola, F. Zollo, and A. Scala, “The COVID-19 gust 16, 2020. social media infodemic,” arXiv preprint arXiv:2003.05004, 2020. [149] “HAMAGEN app - only together will we beat the coronavirus,” https:// [175] N. Karisani and P. Karisani, “Mining coronavirus (COVID-19) posts www.gov.il/en/departments/news/hamagenapp, 2020, Accessed August in social media,” arXiv preprint arXiv:2004.06778, 2020. 16, 2020. [176] J. Samuel, G. Ali, M. Rahman, E. Esawi, Y. Samuel et al., “COVID-19 [150] “An extra tool to fight the epidemic,” https://www.immuni.italia.it, public sentiment insights and Machine Learning for tweets classifica- 2020, Accessed August 16, 2020. tion,” Nawaz and Rahman, Md. Mokhlesur and Esawi, Ek and Samuel, [151] ito, “ito track infections, not people,” https://www.ito-app.org, 2020, Yana, COVID-19 Public Sentiment Insights and Machine Learning for Accessed August 16, 2020. Tweets Classification (April 19, 2020), 2020. [152] https://mask.ir, 2020, Accessed August 16, 2020. [177] D. Prabhakar Kaila, D. A. Prasad et al., “Informational flow on twitter– [153] https://www.mosti.gov.my/web/en/mytrace/, 2020, Accessed August corona virus outbreak–topic modelling approach,” International Jour- 16, 2020. nal of Advanced Research in Engineering and Technology (IJARET), [154] https://www.economie.gouv.fr/stopcovid, 2020, Accessed August 16, vol. 11, no. 3, 2020. 2020. [178] K. Jahanbin and V. Rahmanian, “Using twitter and web news mining [155] https://tracecovid.ae, 2020, Accessed August 16, 2020. to predict COVID-19 outbreak,” Asian Pacific Journal of Tropical [156] “TraceTogether,” https://www.healthhub.sg/apps/38/tracetogether-app, Medicine, vol. 13, 2020. 2020, Accessed August 16, 2020. [179] T. Mackey, V. Purushothaman, J. Li, N. Shah, M. Nali, C. Bardier, B. Liang, M. Cai, and R. Cuomo, “Machine learning to detect self- [157] D. A. Villela, “Discrete time forecasting of epidemics,” Infectious reporting of symptoms, testing access, and recovery associated with Disease Modelling, vol. 5, pp. 189–196, 2020. COVID-19 on twitter: Retrospective big data infoveillance study,” [158] C. Anastassopoulou, L. Russo, A. Tsakris, and C. Siettos, “Data-based JMIR Public Health and Surveillance, vol. 6, no. 2, p. e19509, 2020. analysis, modelling and forecasting of the novel coronavirus (2019- [180] M. K. Siddiqui, R. Morales-Menendez, P. K. Gupta, H. Iqbal, F. Hus- medRxiv nCoV) outbreak,” , 2020. sain, K. Khatoon, and S. Ahmad, “Correlation between temperature and [159] Z. Yang, Z. Zeng, K. Wang, S.-S. Wong, W. Liang, M. Zanin, P. Liu, COVID-19 (suspected, confirmed and death) cases based on Machine X. Cao, Z. Gao, Z. Mai et al., “Modified SEIR and AI prediction Learning analysis,” J Pure Appl Microbiol, vol. 14, 2020. of the epidemics trend of COVID-19 in China under public health [181] A. S. S. Rao and J. A. Vazquez, “Identification of COVID-19 can interventions,” Journal of Thoracic Disease, vol. 12, no. 3, p. 165, be quicker through Artificial Intelligence framework using a mobile 2020. phone–based survey when cities and towns are under quarantine,” [160] D. Liu, L. Clemente, C. Poirier, X. Ding, M. Chinazzi, J. T. Davis, Infection Control & Hospital Epidemiology, vol. 41, no. 7, pp. 826– A. Vespignani, and M. Santillana, “A Machine Learning methodology 830, 2020. for real-time forecasting of the 2019-2020 COVID-19 outbreak using [182] Y. Zhao and H. Xu, “Chinese public attention to COVID-19 epidemic: internet searches, news alerts, and estimates from mechanistic models,” Based on social media,” medRxiv, 2020. arXiv preprint arXiv:2004.04019, 2020. [183] L. Schild, C. Ling, J. Blackburn, G. Stringhini, Y. Zhang, and S. Zan- [161] M. A. Al-Qaness, A. A. Ewees, H. Fan, and M. Abd El Aziz, nettou, “”go eat a bat, chang!”: An early look on the emergence of “Optimization method for forecasting confirmed cases of COVID-19 sinophobic behavior on web communities in the face of COVID-19,” in china,” Journal of Clinical Medicine, vol. 9, no. 3, p. 674, 2020. arXiv preprint arXiv:2004.04046, 2020. [162] M. Abdel-Basset and L. A. Shawky, “Flower pollination algorithm: a [184] M. Gao, H. Igata, A. Takeuchi, K. Sato, and Y. Ikegaya, “Machine comprehensive review,” Artificial Intelligence Review, vol. 52, no. 4, Learning-based prediction of adverse drug effects: an example of pp. 2533–2557, 2019. seizure-inducing compounds,” Journal of Pharmacological Sciences, [163] S. Mirjalili, A. H. Gandomi, S. Z. Mirjalili, S. Saremi, H. Faris, and vol. 133, no. 2, pp. 70–78, 2017. S. M. Mirjalili, “Salp swarm algorithm: A bio-inspired optimizer for [185] R. Lu, X. Zhao, J. Li, P. Niu, B. Yang, H. Wu, W. Wang, H. Song, engineering design problems,” Advances in Engineering Software, vol. B. Huang, N. Zhu et al., “Genomic characterisation and epidemiology 114, pp. 163–191, 2017. of 2019 novel coronavirus: implications for virus origins and receptor [164] D. Karaboga and E. Kaya, “Adaptive network based fuzzy inference binding,” The Lancet, vol. 395, no. 10224, pp. 565–574, 2020. system (ANFIS) training approaches: a comprehensive survey,” Artifi- [186] “The proximal origin of SARS-CoV-2, author=Andersen, Kristian G cial Intelligence Review, vol. 52, no. 4, pp. 2263–2293, 2019. and Rambaut, Andrew and Lipkin, W Ian and Holmes, Edward C and Garry, Robert F, journal=Nature medicine, volume=26, number=4, P. J. Hotez et al., “Vaccine efficacy needed for a covid-19 coronavirus pages=450–452, year=2020, publisher=Nature Publishing Group.” vaccine to prevent or stop an epidemic as the sole intervention,” [187] M. Hoffmann, H. Kleine-Weber, S. Schroeder, N. Kruger,¨ T. Herrler, American journal of preventive medicine, 2020. S. Erichsen, T. S. Schiergens, G. Herrler, N.-H. Wu, A. Nitsche et al., [208] G. A. Weinberg and P. G. Szilagyi, “Vaccine epidemiology: efficacy, “SARS-CoV-2 cell entry depends on ACE2 and TMPRSS2 and is effectiveness, and the translational research roadmap,” The Journal of blocked by a clinically proven protease inhibitor,” Cell, 2020. infectious diseases, vol. 201, no. 11, pp. 1607–1610, 2010. [188] E. Lin and H.-Y. Lane, “Machine Learning and systems genomics [209] A. T. Harding and N. S. Heaton, “Efforts to improve the seasonal approaches for multi-omics data,” Biomarker Research, vol. 5, no. 1, influenza vaccine,” Vaccines, vol. 6, no. 2, p. 19, 2018. p. 2, 2017. [210] W.-H. Chen, U. Strych, P. J. Hotez, and M. E. Bottazzi, “The SARS- [189] L. E. Gralinski and V. D. Menachery, “Return of the coronavirus: 2019- CoV-2 vaccine pipeline: an overview,” Current Tropical Medicine nCoV,” Viruses, vol. 12, no. 2, p. 135, 2020. Reports, pp. 1–4, 2020. [190] Z.-W. Ye, S. Yuan, K.-S. Yuen, S.-Y. Fung, C.-P. Chan, and D.-Y. Jin, [211] N. Restifo, H. Ying, L. Hwang, and W. Leitner, “The promise of nucleic “Zoonotic origins of human coronaviruses,” International Journal of acid vaccines,” Gene Therapy, vol. 7, no. 2, pp. 89–92, 2000. Biological Sciences, vol. 16, no. 10, p. 1686, 2020. [212] WHO, “Draft landscape of COVID-19 candidate vaccines,” [191] J. F.-W. Chan, K.-H. Kok, Z. Zhu, H. Chu, K. K.-W. To, S. Yuan, https://www.who.int/publications/m/item/draft-landscape-of-covid- and K.-Y. Yuen, “Genomic characterization of the 2019 novel human- 19-candidate-vaccines, 2020, Accessed August 04, 2020. pathogenic coronavirus isolated from a patient with atypical pneumonia [213] E. Ong, M. U. Wong, A. Huffman, and Y. He, “COVID-19 coronavirus after visiting wuhan,” Emerging Microbes & Infections, vol. 9, no. 1, vaccine design using reverse vaccinology and Machine Learning,” pp. 221–236, 2020. BioRxiv, 2020. [192] P. K. D. H. John Jumper, Kathryn Tunyasuvunakool [214] G. Liu, B. Carter, T. Bricken, S. Jain, M. Viard, M. Carrington, and and the AlphaFold Team. Computational predictions of D. K. Gifford, “Computationally optimized SARS-CoV-2 MHC class protein structures associated with COVID-19. [Online]. i and ii vaccine formulations predicted to target human haplotype Available: https://deepmind.com/research/open-source/computational- distributions,” Cell Systems, 2020. predictions-of-protein-structures-associated-with-COVID-19 [215] D. Ward, M. Higgins, J. Phelan, M. L. Hibberd, S. Campino, and [193] CDC. Types of coronaviruses. [Online]. Available: https: T. G. Clark, “An integrated in silico immuno-genetic analytical platform //www.cdc.gov/coronavirus/types.html provides insights into COVID-19 serological and vaccine targets,” [194] G. S. Randhawa, M. P. Soltysiak, H. El Roz, C. P. de Souza, K. A. Hill, bioRxiv, 2020. and L. Kari, “Machine Learning using intrinsic genomic signatures for [216] M. S. Rahman, M. N. Hoque, M. R. Islam, S. Akter, A. Rubayet-Ul- rapid classification of novel pathogens: COVID-19 case study,” PLOS Alam, M. A. Siddique, O. Saha, M. M. Rahaman, M. Sultana, K. A. ONE, vol. 15, no. 4, p. e0232391, 2020. Crandall et al., “Epitope-based chimeric peptide vaccine design against [195] G. S. Randhawa, K. A. Hill, and L. Kari, “ML-DSP: Machine Learning S, M and E proteins of SARS-CoV-2 etiologic agent of global pandemic with digital signal processing for ultrafast, accurate, and scalable COVID-19: an in silico approach,” PeerJ, vol. 8, p. e9572, 2020. genome classification at all taxonomic levels,” BMC Genomics, vol. 20, [217] B. Sarkar, M. A. Ullah, F. T. Johora, M. A. Taniya, and Y. Araf, no. 1, p. 267, 2019. “The essential facts of wuhan novel coronavirus outbreak in China and [196] L. Heo and M. Feig, “Modeling of severe acute respiratory syn- Epitope-based vaccine designing against 2019-nCoV,” BioRxiv, 2020. drome coronavirus 2 (SARS-CoV-2) proteins by Machine Learning and [218] M. Prachar, S. Justesen, D. B. Steen-Jensen, S. P. Thorgrimsen, E. Ju- physics-based refinement,” bioRxiv, 2020. rgons, O. Winther, and F. O. Bagger, “COVID-19 vaccine candidates: [197] R. Magar, P. Yadav, and A. B. Farimani, “Potential neutralizing Prediction and validation of 174 SARS-CoV-2 epitopes,” bioRxiv, 2020. antibodies discovered for novel corona virus using Machine Learning,” [219] R. Qiao, N. H. Tran, B. Shan, A. Ghodsi, and M. Li, “Personalized arXiv preprint arXiv:2003.08447, 2020. workflow to identify optimal t-cell epitopes for peptide-based vaccines [198] H. Chen, O. Engkvist, Y. Wang, M. Olivecrona, and T. Blaschke, against COVID-19,” arXiv preprint arXiv:2003.10650, 2020. “The rise of Deep Learning in drug discovery,” Drug discovery today, [220] C. V. Herst, S. Burkholz, J. Sidney, A. Sette, P. E. Harris, S. Massey, vol. 23, no. 6, pp. 1241–1250, 2018. T. Brasel, E. Cunha-Neto, D. S. Rosa, W. C. H. Chao et al., “An [199] B. R. Beck, B. Shin, Y. Choi, S. Park, and K. Kang, “Predicting com- effective ctl peptide vaccine for ebola zaire based on survivors CD8+ mercially available antiviral drugs that may act on the novel coronavirus targeting of a particular nucleocapsid protein epitope with potential (SARS-CoV-2) through a drug-target interaction deep learning model,” implications for COVID-19 vaccine design,” Vaccine, 2020. Computational and Structural Biotechnology Journal, 2020. [221] H. Zhang, K. M. Saravanan, Y. Yang, M. T. Hossain, J. Li, X. Ren, [200] M. Moskal, W. Beker, R. Roszak, E. P. Gajewska, A. Wołos, K. Molga, Y. Pan, and Y. Wei, “Deep Learning based drug screening for novel S. Szymkuc,´ and B. A. Grzybowski, “Suggestions for second-pass coronavirus 2019-nCov,” Interdisciplinary Sciences, Computational anti-COVID-19 drugs based on the Artificial Intelligence measures of Life Sciences, p. 1, 2020. molecular similarity, shape and pharmacophore distribution.” 2020. [222] M. Nielsen, C. Lundegaard, T. Blicher, K. Lamberth, M. Harndahl, [201] S. Jaeger, S. Fulle, and S. Turk, “Mol2vec: Unsupervised Machine S. Justesen, G. Røder, B. Peters, A. Sette, O. Lund et al., “NetMHCpan, Learning approach with chemical intuition,” Journal of chemical in- a method for quantitative predictions of peptide binding to any HLA-A formation and modeling, vol. 58, no. 1, pp. 27–35, 2018. and-B locus protein of known sequence,” PLOS ONE, vol. 2, no. 8, p. [202] O. Kadioglu, M. Saeed, H. Johannes Greten, and T. Efferth, “Iden- e796, 2007. tification of novel compounds against three targets of SARS CoV-2 coronavirus by combined virtual screening and supervised Machine Learning,” Bull. World Health Organ, 2020. [203] N. Bung, S. R. Krishnan, G. Bulusu, and A. Roy, “De novo design of new chemical entities (NCEs) for SARS-CoV-2 using Artificial Intelligence,” ChemRxiv, 2020. [204] A. Zhavoronkov, B. Zagribelnyy, A. Zhebrak, V. Aladinskiy, V. Teren- tiev, Q. Vanhaelen, D. S. Bezrukov, D. Polykovskiy, R. Shayakhmetov, A. Filimonov et al., “Potential non-covalent SARS-CoV-2 3C-like protease inhibitors designed using generative deep learning approaches and reviewed by human medicinal chemist in virtual reality,” 2020. [205] B. Tang, F. He, D. Liu, M. Fang, Z. Wu, and D. Xu, “AI-aided design of novel targeted covalent inhibitors against SARS-CoV-2,” bioRxiv, 2020. [206] F. Hu, J. Jiang, and P. Yin, “Prediction of potential commercially inhibitors against SARS-CoV-2 by multi-task deep model,” arXiv preprint arXiv:2003.00728, 2020. [207] S. M. Bartsch, K. J. O’Shea, M. C. Ferguson, M. E. Bottazzi, P. T. Wedlock, U. Strych, J. A. McKinnell, S. S. Siegmund, S. N. Cox,