Dear Anran, the following recent papers may help you:
Xu J, Glicksberg BS, Su C, et al. Federated Learning for Healthcare Informatics. J Healthc Inform Res 2021;5:1-19. https://link.springer.com/content/pdf/10.1007/s41666-020-00082-4.pdf
Ng D, Lan X, Yao MM, Chan WP, Feng M. Federated learning: a collaborative effort to achieve better medical imaging models for individual sites that have small labelled datasets. Quant Imaging Med Surg 2021;11(2):852-857. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7779924/pdf/qims-11-02-852.pdf
Liu JC, Goetz J, Sen S, Tewari A. Learning From Others Without Sacrificing Privacy: Simulation Comparing Centralized and Federated Machine Learning on Mobile Health Data. JMIR Mhealth Uhealth 2021;9(3):e23728. https://mhealth.jmir.org/2021/3/e23728/PDF
Dou Q, So TY, Jiang M, et al. Federated deep learning for detecting COVID-19 lung abnormalities in CT: a privacy-preserving multinational validation study. NPJ Digit Med 2021;4(1):60. https://www.nature.com/articles/s41746-021-00431-6.pdf
Vaid A, Jaladanki SK, Xu J, et al. Federated Learning of Electronic Health Records to Improve Mortality Prediction in Hospitalized Patients With COVID-19: Machine Learning Approach. JMIR Med Inform 2021;9(1):e24207. https://medinform.jmir.org/2021/1/e24207/PDF
Qian F, Zhang A. The value of federated learning during and post-COVID-19. Int J Qual Health Care. 2021;33(1):mzab010. https://academic.oup.com/intqhc/article/33/1/mzab010/6128587
Federated learning is used guven beavy data sharing restrictions between hospitals. In my opinion the main challenges relate to the distribution mismatch between the different dataset nodes (concept drift, covariate shift, etc). How can we measure this and deal with it?
Saúl Calderón Ramírez for image data, maybe we can measure the pixel distribution? Indeed, there are many kinds of domain-shift. Thus, in my opinion, we may also consider domain adaptation in federated learning.
Yes that could be one way. But perhaps is better to analyze the feature space learned by the models, and do some domain adaptation there as you said. Ive been working in the topic lately. Hopefully later I can share some of the results with you!