Federated Learning for Privacy Protection
Type
Master-Thesis
Title
Federated Learning for Privacy Protection
Supervisor
Abstract
Motivation
Federated learning is expected to break the barriers between data sources while the leakage of data is prevented and was first proposed 2017 by Google. The datasets are distributed across multiple devices and the training of the model can be pushed to the edge.
Goal
- Literature review
- Set up a test infrastructure
Helpful Knowledge
- Fundamentals in security and privacy
- Knowledge in machine learning
- Programming in e.g. Python, C++
Contact
For further information please approach M.Sc. Sascha Löbner (sascha.loebner[at]m-chair.de)
