TheSequence

TheSequence

Share this post

TheSequence
TheSequence
Edge 277: Federated Transfer Learning

Edge 277: Federated Transfer Learning

Federated transfer learning, the TorchFL paper and the OpenFL framework.

Mar 28, 2023
∙ Paid
21

Share this post

TheSequence
TheSequence
Edge 277: Federated Transfer Learning
Share
Created Using Midjourney

In this issue:

  1. An overview of federated transfer learning.

  2. The TorchFL reference architecture for federated learning pipelines.

  3. The OpenFL framework.

💡 ML Concept of the Day: Federated Transfer Learning

In the last two editions of this series, we explored the concepts of horizontal (HFL) and vertical (VFL) federated learning. These two architectures are applicable when nodes in a federation have a high overlap on either the feature or sample space. However, there are many scenarios in which these overlaps are relatively small but still viable for a federated learning scenario. This is where federated transfer learning (FTL) comes into play as a variation of the HFL and VFL architectures that is very applicable in such scenarios.

This post is for paid subscribers

Already a paid subscriber? Sign in
© 2025 Jesus Rodriguez
Privacy ∙ Terms ∙ Collection notice
Start writingGet the app
Substack is the home for great culture

Share