All you need to know about Phybiscus

Phybiscus, a pioneering open-source tool developed within the PAROMA-MED project, offers an innovative approach to Federated Learning (FL).

Available on GitHub: https://github.com/ThalesGroup/pybiscus

Here are six key points to understand Phybiscus and its impact:

1. What is Phybiscus?
   Built on Flower, an open-source Federated Learning framework, Phybiscus provides an easy-to-use pipeline for deploying FL solutions. It is a creation of Thales and is freely available on GitHub: Phybiscus Repository.
2. Why Was Phybiscus Developed?
   As FL frameworks evolve, many remain research-focused, with limitations for practical deployment. Phybiscus addresses these gaps, including eliminating code duplication, simplifying configurations, and separating core components like client, server, and machine learning code.
3. Its Significance in PAROMA-MED
   Phybiscus ensures a robust, easily deployable FL solution, enhancing the project’s testbeds while supporting its mission to integrate reliable, scalable technologies.
4. Sharing With the Community
   By remaining open-source, Phybiscus empowers other projects and researchers to contribute to and benefit from its advancements, fostering innovation across multiple sectors.
5. Benefits
   • Modular: Compatible with any PyTorch model and dataset.
   • Flexible: Uses YAML configurations to streamline deployment without altering core code.
   • User-Friendly: Equipped with a Command Line Interface, detailed debug messages, and containerized versions for seamless use.
   • Code Separation: Ensures deployment ease and scalability.
6. Collaborations and Impact
   Phybiscus is being shared with projects like Secured, solving challenges in creating robust open-source frameworks. It is a vital asset for accelerating progress in Federated Learning while addressing pressing technical and scalability needs.