CoChat
Sign Up
Cited by 156IoT and Edge/Fog ComputingCOVID-19 diagnosis using AIInternet of Things and AI

A survey on deep transfer learning to edge computing for mitigating the COVID-19 pandemic

Open access
Sufian, Abu, Ghosh, Anirudha, Sadiq, Ali Safaa, Smarandache, Florentin·2020·Journal of Systems Architecture
DOI 10.1016/j.sysarc.2020.101830PMC PMC7326453

Published version of a preprint

Open this paper in CoChat — free account

Abstract

Global Health sometimes faces pandemics as are currently facing COVID-19 disease. The spreading and infection factors of this disease are very high. A huge number of people from most of the countries are infected within six months from its first report of appearance and it keeps spreading. The required systems are not ready up to some stages for any pandemic; therefore, mitigation with existing capacity becomes necessary. On the other hand, modern-era largely depends on Artificial Intelligence(AI) including Data Science; and Deep Learning(DL) is one of the current flag-bearer of these techniques. It could use to mitigate COVID-19 like pandemics in terms of stop spread, diagnosis of the disease, drug & vaccine discovery, treatment, patient care, and many more. But this DL requires large datasets as well as powerful computing resources. A shortage of reliable datasets of a running pandemic is a common phenomenon. So, Deep Transfer Learning(DTL) would be effective as it learns from one task and could work on another task. In addition, Edge Devices(ED) such as IoT, Webcam, Drone, Intelligent Medical Equipment, Robot, etc. are very useful in a pandemic situation. These types of equipment make the infrastructures sophisticated and automated which helps to cope with an outbreak. But these are equipped with low computing resources, so, applying DL is also a bit challenging; therefore, DTL also would be effective there. This article scholarly studies the potentiality and challenges of these issues. It has described relevant technical backgrounds and reviews of the related recent state-of-the-art. This article also draws a pipeline of DTL over Edge Computing as a future scope to assist the mitigation of any pandemic.

AI snapshot

Written from the paper’s full text.

TL;DR

This survey reviews deep learning, deep transfer learning, edge computing, and datasets for mitigating COVID-19-like pandemics, and proposes a conceptual DTL-over-EC pipeline for future outbreak mitigation. It reports that DTL can reduce data/compute demands and EC can lower latency/privacy issues, but notes limited peer-reviewed COVID-19 studies and no experimental validation of the proposed DTL-EC model.

Key findings

  • The survey identifies three main DL challenges for pandemics: shortage of reliable datasets, geographic/environmental data variation due to virus mutation, and high computational resource requirements; DTL/few-shot learning/GAN and EC are suggested mitigations.
  • DTL studies for COVID-19 mostly focus on chest image analysis; examples include Minaee et al. achieving sensitivity 97% (5%) and specificity ~90% on chest X-rays, Basu et al. ~95.3% overall accuracy, and Cohen et al. severity score MAE 1.14 and lung opacity MAE 0.78.
  • DL screening examples include Wang et al. CT-based CNN with 89.5% accuracy, 87% sensitivity, and 88% specificity, and L. Wang et al. chest X-ray screening with 83.5% accuracy.
  • The proposed DTL-EC pipeline would train a DL model on a cloud with benchmark data, push the pre-trained model to edge devices, and fine-tune last layers locally for tasks such as ICU monitoring, patient care, and social distancing/PPE monitoring; the paper says this idea has not been studied or experimented for COVID-19.

Methods

  • Design: narrative survey/systematic study of DL, DTL, EC, and datasets for pandemic mitigation, with tables of recent works and a proposed conceptual DTL-EC pipeline.
  • Data/sample: literature and datasets; includes tables of DL works, DTL works, EC works, and COVID-19-related datasets (e.g., COVID-CT-Dataset 275 positives; COVIDx CXR 13,800 images across 13,725 patients; chest CT 4,356 exams from 3,322 patients).
  • Analysis: qualitative synthesis of potentialities and challenges, with no new experiments or quantitative meta-analysis reported.

Limitations

  • Paper states: limited peer-reviewed studies and experimental results because the pandemic is new, so preprints were included; data insufficiency, regional variation, privacy/security, and limited EC pandemic studies remain challenges.
  • Paper states the proposed DTL-EC model is conceptual and “has not been studied or experimented to mitigate COVID-19 pandemic”; future challenges include sensor networking, model selection, EC-Fog-Cloud integration, security/privacy, and simulation.
  • Evident from the text: no empirical validation, benchmark comparison, or formal systematic-review protocol is reported for the proposed pipeline; tables are illustrative rather than exhaustive.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1016/j.sysarc.2020.101830

Cites

158 of 158 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 38 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    Only a few among the existing DTL studies are proposed for target drug interaction, cough sound classification, and similar tasks.

    “Only a few among them are proposed for target drug interaction, cough sound classification, etc.”
    Quote in the paperNot a substantial claim
  • Findingown data

    Most existing studies and experiments on COVID-19 were applied for chest image analysis.

    “The number of studies is not many; in addition most of the existing studies and experiments on COVID-19, were applied for chest image analysis as reported in Table 2.”
    Quote in the paperNot a substantial claim
  • Findingown data

    From Table 1, the majority of the DL works are for assisting radiologists to diagnose diseases.

    “From Table 1 it could be drawn one conclusion that the majority of the works are for assisting radiologists to diagnose diseases.”
    Quote in the paperNot a substantial claim
  • Premisecitation

    The requirement of a large amount of labeled data is a big problem in solving critical domain-specific tasks, specifically medical applications, where making large-scale, high-quality annotated medical datasets is very complex and expensive.

    “So, the requirement of a large amount of labeled data is a big problem in solving some critical domain-specific tasks, specifically the applications for the medical domain, where the making of large-scale, high-quality annotated medical datasets is very complex, and expensive [40].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [40] Altman, 2017, Artificial intelligence (ai) systems for interpreting complex medical datasets, Clin. Pharmacol. Therapeut. source

      Not checked in the free run.

  • Premisecitation

    The usual DL model requires large computing power such as a GPU-enabled server.

    “In addition, the usual DL model requires large computing power such as GPU enabled sever, although researchers are trying hard to optimizing it [41], [42].”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [41] LeCun, 2019, 1.1 deep learning hardware: past, present, and future

      Not checked in the free run.

    • cites [42] Mittal, 2019, A survey of techniques for optimizing deep learning on gpus, J. Syst. Architect. source

      Not checked in the free run.

  • May combine two claims. 23
  • Classified as probably not a substantial claim. 32

  • May combine two claims.

    Findingnone given

    So far only a limited number of studies have investigated the use of Edge Computing to obtain an efficient and effective mitigation system of COVID-19, and only a few EC studies in pandemic are proposed in limited areas of application.

    “So far only a limited number of studies have investigated the use of EC in obtain an efficient and effective mitigation system of COVID-19.”
    • Also stated in: § Edge Computing to Mitigate Pandemic · plain
      “But so far only a few studies on EC in pandemic are proposed in limited areas of application as mentioned in Table 3; this table also mentioned some non-pandemic but related works.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    DTL in edge devices as an Edge Computing approach could be a smart technique to mitigate a new pandemic.

    “Therefore, DTL in edge devices as an EC could be smart techniques to mitigate a new pandemic [17].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [17] Wang, 2020, Convergence of edge computing and deep learning: acomprehensive survey, IEEE Commun. Surv. Tutor. source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Edge or IoT devices such as IoT, webcam, drone, intelligent medical equipment, robot, and wearable sensors are very useful in any pandemic situation and make infrastructures sophisticated and automated, helping to cope with outbreaks.

    “On the other hand edge devices such as IoT, Webcam, Drone, Intelligent Medical Equipment, Robot, etc. are very useful in any pandemic situation. These types of equipment make the infrastructures sophisticated and automated which helps to cope with an outbreaks [15].”
    • Also stated in: § Edge Computing to Mitigate Pandemic · plain
      “Edge or IoT devices-based sophisticated equipments such as smart medical equipment, webcam, drone, wearable sensors, etc. are very useful in a pandemic like situations [112].”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [15] Ting, 2020, Digital technology and covid-19, Nature Med.

      Not checked in the free run.

    • cites [112] Zhu, 2020, IoT pcr for pandemic disease detection and its spread monitoring, Sensor. Actuator. B source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Appropriate drugs, vaccines, and infrastructure are not available up to some stages of any outbreak, so mitigating such diseases with existing capacity is most important in those stages.

    “The appropriate drugs, vaccines, infrastructure, etc. are not available up to some stages of any outbreaks. Therefore, mitigate these types of diseases with existing capacity becomes most important in those stages [7], [8].”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [7] Organization, 2020, Rational use of personal protective equipment for coronavirus disease (COVID-19) and considerations during severe shortages: interim guidance, 6 April 2020

      Not checked in the free run.

    • cites [8] P. Daszak, K.J. Olival, H. Li, A strategy to prevent future pandemics similar to the 2019-ncov outbreak, 2020. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    DL methods have been successfully applied for several AI-based medical applications such as MRI image analysis for cancer and diabetes diagnoses, and conjunction with biometric characteristics.

    “Nowadays, DL methods have been successfully applied for several AI-based medical applications such as Magnetic Resonance Imaging (MRI) images analysis for cancer and diabetes diagnoses, conjunction with biometric characteristics, etc [20].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [20] Litjens, 2017, A survey on deep learning in medical image analysis, Med. Image Anal. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    COVID-19 is a disease caused by a novel coronavirus called SARS-CoV-2, and the virus is transferable from human to human with high infection factors.

    “The COVID-19 is a disease caused by a novel coronavirus called ‘SARS-CoV-2’. This virus is transferable from human to human and it’s spreading, and infection factors are very high [1], [2].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [1] Rothan, 2020, The epidemiology and pathogenesis of coronavirus disease (covid-19) outbreak, J. Autoimmun. source

      Not checked in the free run.

    • cites [2] Jin, 2020, Virology, epidemiology, pathogenesis, and control of covid-19, Viruses source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    DL techniques could assist to mitigate COVID-19-like pandemics in terms of stopping spread, diagnosis of the disease, drug and vaccine discovery, treatment, patient care, and more.

    “Therefore, these techniques could also assist to mitigate COVID-19 like pandemics in terms of stop spread, diagnosis of the disease, drug & vaccine discovery, treatment, patient care, and many more [12], [13] 132].”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [12] Shi, 2020, Review of artificial intelligence techniques in imaging data acquisition, segmentation and diagnosis for covid-19, arXiv preprint arXiv:2004.02731

      Not checked in the free run.

    • cites [13] A. Sufian, D.S. Jat, A. Banerjee, Insights of artificial intelligence to stop spread of covid-19, in: Big Data Analytics and Artificial Intelligence against COVID-19: Inovative Vision and Approach.

      Not checked in the free run.

  • May combine two claims.

    Interpretationcitation

    For a new pandemic, data insufficiency and variation over different geographic regions is a huge problem, so Deep Transfer Learning would be effective because it learns from one task and can apply to another task after required fine-tuning.

    “For a new pandemic, data insufficiency and it’s variation over different geographic regions is a huge problem, so here, Deep Transfer Learning (DTL) would be effective as it learns from one task and could apply in another task after required fine-tuning [14].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [14] Tan, 2018, A survey on deep transfer learning

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    DL is being used or could be used for purposes such as patient care and detection of systematic social distancing violation.

    “In addition, DL is using or could be used for other purposes such as patient care, detection systematic social distancing violation, etc [13].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [13] A. Sufian, D.S. Jat, A. Banerjee, Insights of artificial intelligence to stop spread of covid-19, in: Big Data Analytics and Artificial Intelligence against COVID-19: Inovative Vision and Approach.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    DL can make more accurate classifications and predictions with the help of innovative algorithms, computing power of modern machines, and the availability of Big Datasets.

    “In addition, DL can make more accurate classifications and predictions with the help of innovative algorithms, computing power of modern machines, and the availability of Big Datasets [19].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [19] Goodfellow, 2016

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given

    The novel coronavirus behaves differently across geographic regions and different environments.

    “As said this novel coronavirus is behaving differently across geographic regions, different environments, etc.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    The global standard diagnostic pathogenic laboratory testing is massively time-consuming and costly, with significant false-negative results.

    “But the global standard diagnostic pathogenic laboratory testing is massive time consuming and more costly with significant false-negative results [51].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [51] Santosh, 2020, A review of salivary diagnostics and its potential implication in detection of covid-19, Cureus

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Edge Computing focuses on starting the computational process close to IoT devices or the Edge in order to reduce utilized bandwidth and latency.

    “Therefore, instead of relying on doing all the work at a cloud, it focuses to start the computational process close to the IoT devices or Edge (near to the source of data) in order to reduce the utilized bandwidth and latency [55], [56].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [55] Satyanarayanan, 2017, The emergence of edge computing, Computer source

      Not checked in the free run.

    • cites [56] Garca-Valls, 2018, Introducing the new paradigm of social dispersed computing: applications, technologies and challenges, J. Syst. Architect. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

    Interpretationreasoning

    The combined DTL-EC model will take the power of DL through DTL and would be applicable in critical sectors by EC to cope with a sudden pandemic.

    “This combined model, that is, Deep Transfer Learning over Edge Computing(DTL-EC) will take the power of DL through DTL as well as would be applicable in critical sectors by EC to cope with a sudden pandemic.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Data privacy and security are big issues.
    • Rests on an assumption: Edge devices are capable of performing fine-tuning of deep learning models.
    • Rests on an assumption: A small task-oriented dataset is sufficient for fine-tuning a pre-trained model at the edge to achieve required inference.

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

    Premisenone given

    Data is the main driving force that brings knowledge, but it is not easily available.

    “As mentioned data is the main driving force to which bring the knowledge but it not easily available.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationcitation

    Most of the potential COVID-19 mitigation works could be easier when AI is cooperating and forming such a model along with IoT or Edge Devices.

    “Most of these works could be easier when AI is cooperating and forming such a model along with IoT or ED [13].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [13] A. Sufian, D.S. Jat, A. Banerjee, Insights of artificial intelligence to stop spread of covid-19, in: Big Data Analytics and Artificial Intelligence against COVID-19: Inovative Vision and Approach.

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisereasoning

    Deep transfer learning methods provide task adaptability where data shortages exist.

    “Section 2.2 has described the task adaptability through DTL methods where data shortages are there.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisenone given

    Edge computing has potential where computing power is low.

    “Section 2.3 mentioned the potentialities of EC where computing power is low.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationcitation

    A main challenge is the variety of data of a pandemic virus; SARS-CoV-2 mutates over different geographic regions, environments, and time, so a pandemic dataset collected from one region may not work to draw inference on the pandemic of other regions.

    “The second one is the variety of data of a pandemic virus. This COVID-19 virus ‘SARS-CoV-2’ has been mutating itself over different geographic regions, environments, and time [102], [103]. Therefore, the pandemic dataset collected from one region may not be work to drawn inference on the pandemic of other regions.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [102] Shen, 2020, Genomic diversity of sars-cov-2 in coronavirus disease 2019 patients, Clin. Infect. Dis. source

      Not checked in the free run.

    • cites [103] Grubaugh, 2020, We shouldn’t worry when a virus mutates during disease outbreaks, Nature Microbiol. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Recent AI techniques mostly follow data-driven approaches.

    “Recent AI techniques are mostly follow data driven approaches [122], [123].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [122] Zhang, 2018, Data driven feature selection for machine learning algorithms in computer vision, IEEE Internet Things J. source

      Not checked in the free run.

    • cites [123] Li, 2020, A survey of data-driven and knowledge-aware explainable ai, IEEE Trans. Know. Data Eng. source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisenone given

    Most of the datasets mentioned in Table 4 are for clinical purposes.

    “Although some datasets mentioned in Table 4 but most of them are for clinical purposes.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing). May combine two claims.

    Interpretationnone given

    Edge Computing works on site, so many benefits could be drawn from Edge Computing with IoT or Edge Devices.

    “The EC works on site, so, many benefits could draw from EC with IoT or ED.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given

    DTL does task adaptation that is very necessary for analyzing, diagnosing, and mitigating COVID-19-like pandemics.

    “The DTL does task adaption that is very necessary for analyzing, diagnosing as well as mitigating COVID-19 like pandemics.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

    Premisenone given

    For COVID-19, many sectors and organizations are very active, and as a result many data sources are quickly oriented towards the COVID-19 pandemic.

    “Although for COVID-19, many sectors, organizations are very active as a result many data sources are quickly oriented towards COVID-19 pandemic.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given

    Deep learning has limitations in coping with the challenges of a pandemic.

    “the DL has some limitations to cope with the challenges of a pandemic”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    To cope with a pandemic, data is one of the driving forces.

    “Therefore, to cope with a pandemic, data is one of the driving forces.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    One possibility of achieving a better system is when DTL is implemented alongside Edge Computing.

    “One possibility of archiving this when DTL implemented alongside with EC which as conceptually describes in Section 5.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationcitation

    Edge computing brings computation near edge devices and reduces latency, security and privacy issues, so this computing paradigm will be very effective to mitigate a pandemic situation.

    “As mentioned in Section 2.3, edge computing brings the computation to near edge devices. It reduces latency, security & privacy issue, etc. Therefore, this computing paradigm will be very effective to mitigate a pandemic situation [113].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [113] Al-Zinati, 2019, An agent-based self-organizing model for large-scale biosurveillance systems using mobile edge computing, Simulat. Modell. Pract. Theory source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based).

    Premisereasoning

    DL or DTL based algorithms almost fully depend on the dataset.

    “DL or DTL based algorithms almost fully depend on the dataset.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Interpretationreasoning

    For Cloud or even Fog Computing, latency and data security and privacy could be a problem; therefore Edge Computing could be effective for the third challenge.

    “However, for Cloud or even Fog Computing latency and data security & privacy could be a problem. Therefore, Edge Computing could be effective for the third challenge, which has described in Section 4.3.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: For Cloud or even Fog Computing, latency and data security and privacy could be a problem.

  • May combine two claims.

    Interpretationcitation

    A main challenge is the high computational resources required for a DL-based model, whereas IoT or Edge Devices useful for many purposes have low computing resources.

    “The third one high computational resources required for a DL-based model whereas to cope with an outbreak IoT or Edge Device (ED) are useful for many purposes [15]. Though these types of equipments have low computing resources.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [15] Ting, 2020, Digital technology and covid-19, Nature Med.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationcitation

    For the first two challenges of applying DL to pandemics, DTL or few-shot learning and GAN could be a possible approach towards possible solutions.

    “For the first two challenges, DTL or few shot learning and GAN [28] could be a possible approach towards possible solutions.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [28] Goodfellow, 2014, Generative adversarial nets

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationnone given

    A better system could be delivered when the most suitable algorithm is applied on Edge Computing.

    “Though a better system could be delivered when the most suitable algorithm applied on EC.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationcitation

    DTL could be effective to mitigate COVID-19-like pandemics because sufficient datasets are difficult to develop in a short period, so a DL model could be trained using a large-scale benchmark or available dataset and learned features could be used in the COVID-19 domain.

    “Therefore, the DTL could be effective in this case. As through DTL, a DL model could be trained using a large scale benchmark or available dataset and learned features could be used in the domain of COVID-19 [53].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [53] Apostolopoulos, 2020, Covid-19: automatic detection from x-ray images utilizing transfer learning with convolutional neural networks, Phys. Eng. Sci. Med.

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Premisereasoning

    For COVID-19 or a sudden pandemic or epidemic, gathering data and arranging it in a knowledgeable form is not an easy task.

    “Specially COVID-19 or a sudden pandemic or epidemic, gathering data and arrange it in a knowledgeable form are not expected as an easy task.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationreasoning

    The model may be successfully deployed in critical sectors such as hospitals, airports, markets, emergency service areas, and primary hotspots for spreading pandemics.

    “This model may be successfully deployed in some critical sectors such as hospitals, airports, markets, emergency service areas, and those areas which are the primary hotshots for spreading pandemics.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Edge devices are capable of performing fine-tuning of deep learning models.

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

    Interpretationreasoning

    The DTL-EC model could be helpful in the healthcare sector, quarantine centers, or other critical areas where an outbreak may arise, and may be used for remote health monitoring and elderly care.

    “DTL-EC model could be helpful in the healthcare sector, quarantine center, or other critical areas where an outbreak may arise as well as it may be used for remote health monitoring, elderly care, etc.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Edge devices are capable of performing fine-tuning of deep learning models.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationnone given

    The running pandemic shall be mitigated but there will be a left over impact on global health, economics, education, etc., so mitigation of this pandemic is necessary to restrict further worsening.

    “The running pandemic shall be mitigated but there will be a left over impact on global health, economics, education, etc, so mitigation of this pandemic is necessary to restrict further worsen.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Some studies exist in DTL-EC, as in [68], and related work is mentioned in Section 3.

    “There are some studies that exist in DTL-EC as in [68] and some related work mentioned in Section 3.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [68] Li, 2018, Learning IoT in edge: Deep learning for the internet of things with edge computing, IEEE Netw. source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisereasoning

    There is a limited number of peer-reviewed studies and experimental results on the running pandemic.

    “As the running pandemic is very new, so, there is a limited number of peer-reviewed studies and experimental results.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationnone given

    The model could also be used for remote health monitoring or elderly care.

    “It could also be used for remote health monitoring or elderly care.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Edge devices are capable of performing fine-tuning of deep learning models.

  • May combine two claims.

    Premisenone given

    Existing DTL-EC works are still in general concept or their proposed methods are applicable only to some other application areas.

    “However, these works are still in general concept or their proposed methods are applicable only to some others application areas.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

    Interpretationnone given

    For the transfer learning approach, only edge computing is not sufficient; adoption of an EC-Fog-Cloud combined model would be more useful.

    “For the transfer learning approach, only EC is not sufficient, while the adoption of EC-Fog-Cloud combined model would be more useful.”
    Quote in the paperNot a substantial claim
43 of 43

  • The paper relies on this without support.

    Tests are hardly able to take place in common healthcare centres or hospitals due to limited resources and places compared with the high volume of cases at one time.

    “At the same time, tests are are hardly to be taken place in the common healthcare centres or hospital due to limited resources and places compared with the high volume of cases at one time.”

  • The paper relies on this without support.

    For Cloud or even Fog Computing, latency and data security and privacy could be a problem.

    “However, for Cloud or even Fog Computing latency and data security & privacy could be a problem.”

  • The paper relies on this without support.

    Data privacy and security are big issues.

    “Data privacy and security also are considered ones of the big issues.”

  • The paper relies on this without support.

    Edge devices are capable of performing fine-tuning of deep learning models.

    “In edge devices required fine-tuning mechanism to be implemented into that model with some real data with ground truth.”

  • The paper relies on this without support.

    A small task-oriented dataset is sufficient for fine-tuning a pre-trained model at the edge to achieve required inference.

    “So, at least a small task oriented dataset needs to be created.”
5 of 5

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Reference 65 is assigned to two different studies

    Minor

    The text attributes an In-Edge AI federated-learning framework by X. Wang and colleagues to reference 65, but reference 65 is a different paper on edge-server placement by S. Wang and colleagues.

  • Reported binary-classification accuracy exceeds both sensitivity and specificity

    Minor

    An accuracy of 0.996 is 99.6%. For the same binary classification, accuracy must lie between sensitivity and specificity, but both reported values are lower.

2 of 2

Cites

1 retracted

Flagged first, then by citations

Research works printed but not linked: 1

  1. Li, 2020, Artificial intelligence distinguishes covid-19 from community acquired pneumonia on chest ct, Radiology, 200905

Cited by

Sorted by citations

Related

Authors
Sufian, A., Ghosh, A., Sadiq, A. S., & Smarandache, F.
Affiliations
  • University of Gour Banga (IN)
  • University of Wolverhampton (GB)
Corresponding author
Sufian, Abu
ORCID
Venue
Journal of Systems Architecture
Publisher
Elsevier BV
Published
2020-09
Volume
108
Pages
101830
ISSN
1383-7621 (ISSN-L)
DOI
10.1016/j.sysarc.2020.101830
PMC
PMC7326453
OpenAlex
W3037096201
Cited by
156
References
158
Citations by year
2026: 5 · 2025: 13 · 2024: 11 · 2023: 27 · 2022: 46 · 2021: 44 · 2020: 10
FWCI
14.90
Type
journal article
Language
en
Open access
Licence
Locations
  • Journal of Systems Architecture (publishedVersion)
  • PubMed Central (publishedVersion)
  • Wolverhampton Intellectual Repository and E-Theses (University of Wolverhampton) (acceptedVersion, cc-by-nc-nd)
  • Figshare (submittedVersion, cc-by)
Topics
  • Physical Sciences › Computer Science › Computer Networks and Communications › IoT and Edge/Fog Computing
  • Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › COVID-19 diagnosis using AI
  • Physical Sciences › Computer Science › Information Systems › Internet of Things and AI
Keywords
deep transfer learning, edge computing, COVID-19 mitigation, pandemic response, edge devices, resource-constrained devices, Internet of Things, diagnosis, data scarcity, resource-constrained computing

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-09.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-09.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review