Publication detail

A Proposal for a Federated Learning Protocol for Mobile and Management Systems

MICHÁLEK, J. OUJEZSKÝ, V. HOLÍK, M. ŠKORPIL, V.

Original Title

A Proposal for a Federated Learning Protocol for Mobile and Management Systems

Type

journal article in Web of Science

Language

English

Original Abstract

In this research paper, we introduce a federated learning communication protocol tailored for emergency management applications. Our primary objective is to tackle the communication challenges that arise in such critical scenarios. In order to overcome the limitations associated with centralized server architectures, we present an innovative communication protocol. This protocol empowers the framework to effectively cooperate with multiple centralized servers, fostering efficient knowledge sharing and model training while ensuring the utmost data privacy and security. By harnessing this protocol, our framework elevates the performance and resilience of vital infrastructure systems operating on the Android platform, thereby facilitating real-time operational scenarios. This research makes a substantial contribution to the field of emergency management applications, as we offer a comprehensive solution that optimizes communication and enables seamless collaboration with numerous centralized servers.

Keywords

Android; communication protocol; federated learning; framework; machine learning; mobile

Authors

MICHÁLEK, J.; OUJEZSKÝ, V.; HOLÍK, M.; ŠKORPIL, V.

Released

21. 12. 2023

Publisher

MDPI

ISBN

2076-3417

Periodical

Applied Sciences - Basel

Year of study

14

Number

1

State

Swiss Confederation

Pages from

1

Pages to

14

Pages count

14

URL

Full text in the Digital Library

BibTex

@article{BUT186768,
  author="Jakub {Michálek} and Václav {Oujezský} and Martin {Holík} and Vladislav {Škorpil}",
  title="A Proposal for a Federated Learning Protocol for Mobile and Management Systems",
  journal="Applied Sciences - Basel",
  year="2023",
  volume="14",
  number="1",
  pages="1--14",
  doi="10.3390/app14010101",
  issn="2076-3417",
  url="https://www.mdpi.com/2076-3417/14/1/101"
}