A Collaborative Framework for Facilitating Federated Learning among Jupyter Users
Anandan Krishnasamy, Yuandou Wang, Zhiming Zhao · 2024
Federated learning (FL) allows multiple partners to train machine learning models without sharing raw data, thus preserving privacy. Despite its promising aspects, existing FL frameworks have some drawbacks regarding flexibility, decentralized aggregation, and collaborative environments. This poster presents FedLearn, a collaborative community framework built atop the JupyterLab environment for FL among Jupyter users. We use a microservices architecture to implement the framework and enable automated FL deployment processes across multiple clouds. The demonstration showcases the feasibility of the FedLearn community portal for Jupyter users.