FLPanGuard a Conceptual FL-Based Framework to Detect and Fight New Pandemics

Case of COVID-19 Pandemic

Abstract

The COVID-19 pandemic has highlighted the importance of artificial intelligence (AI) and data-driven solutions in enabling rapid response and decision-making support. To contain the pandemic, medical professionals must collaborate with experts in other fields, such as data scientists, to analyze medical data, including electronic health records (EHR), and to find solutions as fast as possible. However, within a sensitive industry such as healthcare, significant challenges related to data access persist due to privacy concerns and regulations. To solve these challenges, federated learning (FL) provides a novel approach to pandemic response by enabling the use of decentralized private data while maintaining data privacy. This paper provides an overview of FL and explores its application in healthcare during the COVID-19 pandemic. It also introduces FLPanGuard, a conceptual framework that focuses on detecting and fighting a pandemic using FL and other technologies to enhance data collection and security. This framework provides a global response to pandemics starting in the pre-pandemic phase by predicting potential outbreaks using social media content and EHR. It is also expanding into the pandemic phase by assisting medical professionals in diagnosis and treatment. Finally, we explore the limitations of FL to highlight areas where further research is required.