The World Health Organization (WHO) proclaimed the coronavirus of 2019 (COVID-19) a global pandemic in March 2020. Effective testing is essential to stop the epidemic from spreading. Using various clinical symptom data, artificial intelligence and machine learning approaches can aid in COVID-19 detection. While the federated learning (FL) strategy depending on decentralized data helps safeguard data privacy, a crucial factor in the health domain, the deep learning (DL) approach requiring centralized data is exposed to a significant risk of data privacy breaches. With an emphasis on both, this chapter examines current developments in using DL and FL approaches for COVID-19 identification.
Top1. Introduction
COVID-19, short for “Coronavirus Disease 2019,” is a highly contagious respiratory illness caused by the novel coronavirus SARS-CoV-2 (Alhejily. 2021) (Barua et al., 2023). It first emerged in late 2019 in Wuhan, China, and has since led to a global pandemic (Matsushita et al., 2021). The virus can cause a range of symptoms, from mild respiratory issues to severe pneumonia, and it has had significant social, economic, and healthcare impacts worldwide (Faroux et al., 2020) (Barua et al., 2021). Preventative measures like vaccination, mask-wearing, and social distancing have been crucial in controlling its spread. Efforts to combat COVID-19 continue through vaccination campaigns, research into treatments, and ongoing public health measures to prevent transmission (Singh et al., 2021) (Liu et al., 2022). It's essential to stay informed about the latest guidance from health authorities to protect yourself and others from the virus. Despite early preventative measures, high-quality clinical measures, and the required application of public health practices (Barua et al., 2022), coronavirus infections are still skyrocketing internationally even if different regions of the world are at different stages of outbreak (Nepali et al., 2022). Everyone feels a growing sense of urgency to contain the COVID-19 spread by effective testing and isolation. The research community can contribute by utilizing cutting-edge artificial intelligence (AI) techniques to produce fresh ideas and approaches for COVID-19 detection (Barua et al., 2022) (Mostafa et al., 2021) (Ezugwu et al., 2021). This is made possible by the fact that the number of COVID-19 cases has significantly increased, allowing for the daily collection of a sizable amount of pertinent data. Machine learning (ML) helps address COVID-19 concerns by improving diagnosing capability (Schrempft et al., 2023) (Barua et a., 2022), modelling methodologies (Alyasseri et al., 2022), and forecasting potential epidemics thanks to breakthroughs in computer technologies, access to massive data, and major algorithmic developments (van der Schaar et al., 2021) (Barua et al., 2023). Traditional machine learning (ML), as depicted in Figures 1, relies on manually extracted features, which are not only prone to error but also time-consuming and laborious to produce, especially in conditions similar to COVID-19, where data is extremely sensitive and widely dispersed.
Figure 1. Conventional machine learning