Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Flotta: a Secure and Flexible Spark-inspired Federated Learning Framework

Claudio Bonesana, Daniele Malpetti, Sandra Mitrović, Francesca Mangili, Laura Azzimonti

TL;DR

Flotta is a Federated Learning framework designed to train machine learning models on sensitive data distributed across a multi-party consortium conducting research in contexts requiring high levels of security, such as the biomedical field.

Abstract

We present Flotta, a Federated Learning framework designed to train machine learning models on sensitive data distributed across a multi-party consortium conducting research in contexts requiring high levels of security, such as the biomedical field. Flotta is a Python package, inspired in several aspects by Apache Spark, which provides both flexibility and security and allows conducting research using solely machines internal to the consortium. In this paper, we describe the main components of the framework together with a practical use case to illustrate the framework's capabilities and highlight its security, flexibility and user-friendliness.

Flotta: a Secure and Flexible Spark-inspired Federated Learning Framework

TL;DR

Flotta is a Federated Learning framework designed to train machine learning models on sensitive data distributed across a multi-party consortium conducting research in contexts requiring high levels of security, such as the biomedical field.

Abstract

We present Flotta, a Federated Learning framework designed to train machine learning models on sensitive data distributed across a multi-party consortium conducting research in contexts requiring high levels of security, such as the biomedical field. Flotta is a Python package, inspired in several aspects by Apache Spark, which provides both flexibility and security and allows conducting research using solely machines internal to the consortium. In this paper, we describe the main components of the framework together with a practical use case to illustrate the framework's capabilities and highlight its security, flexibility and user-friendliness.
Paper Structure (12 sections, 2 figures)

This paper contains 12 sections, 2 figures.

Table of Contents

  1. Introduction
  2. Background
  3. FL frameworks
  4. Yet another framework
  5. Ferdelance
  6. Workbench
  7. Nodes
  8. Artifacts
  9. How to use Ferdelance
  10. Node side
  11. Workbench side
  12. Conclusions

Figures (2)

  • Figure 1: [Not a final version of course, symbols can change and the style as well. This is just for proposing the content.]
  • Figure 2: [The idea is to update this old figure by representing party one with both the workbench and a data node, and clearly showing that the workbench is not a node. I would have "Party 1" as a tag, and inside the node, and attached to the node the workbench. The numbers will slightly change too, as we need to introduce a number for local training and one for communicating results (see description at the beginning of section III).]