Authors: Hakan Uzuner
Standard machine learning approaches require a huge amount of training data to be stored centralized in order to feed the learning algorithms. Keeping and using data centralized brings many negative aspects with it. Those aspects can be inefficient communication between the centralized data center and the clients producing the data, privacy issues and quick usability of the profits and results of the training. Google’s new approach, federated learning, on the other hand tackles all these problems. The training data is kept decentralized at the client’s devices while communicating only with small updates of the common model. This method allows for optimizations of communication, keeping the privacy of users involved in the process and providing quick usability of the model’s process. In this paper I will explain how the federated learning principle works. Further on, I will give a small insight on optimization possibilities of communication efficiency as well as on privacy issues involved in machine learning processes and how those can be solved using federated learning principles. Additionally, I will show the connection between the federated learning concept and organic computing.
Comments: 5 Pages.
Download: PDF
[v1] 2019-03-10 05:27:02
Unique-IP document downloads: 78 times
Vixra.org is a pre-print repository rather than a journal. Articles hosted may not yet have been verified by peer-review and should be treated as preliminary. In particular, anything that appears to include financial or legal advice or proposed medical treatments should be treated with due caution. Vixra.org will not be responsible for any consequences of actions that result from any form of use of any documents on this website.
Add your own feedback and questions here:
You are equally welcome to be positive or negative about any paper but please be polite. If you are being critical you must mention at least one specific error, otherwise your comment will be deleted as unhelpful.