A Systematic Analysis on Active Federated Learning Network in Healthcare Industry

In recent years, many researchers have demonstrated that privacy acumen can be applied to healthcare applications to enhance patients’ life quality using Machine Learning (ML) and Deep Learning (DL). As far as we are aware, a tiny degree of patient privacy is contained in medical data, which necessitates extreme data protection against disclosure to intruders. In machine learning and deep learning, to build accurate system data pooling done at server, causes data breach, to overcome this issue, in 2016, Google proposed an idea to share learned models rather than sharing the data called Federated Learning (FL). In this systematic analysis, we lay out FL's primary objective concerning privacy. Afterwards, we presented the primary challenges in configuring and deploying federated learning networks and introduce algorithms within advanced federated learning architectures by integrating Convolutional Neural Networks (CNNs), a blockchain-based mechanism, and a non-fungible token (NFT) for the detection of various harmful diseases. moreover, we analysed the configuration settings, performances, Limitations of different algorithms against different datasets. Ultimately, we conclude this analysis by discussing optimization and communication challenges in future FL smart healthcare systems. We believe that reading this through can be beneficial for academicians, industry experts, and neoteric alike, providing them with direction and suggestion for their next endeavours.