Online Federated Learning Applied to Cybersecurity

• Grant number: 24/18548-7
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: December 01, 2024
• End date: November 30, 2025
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal Investigator: Daniel Macêdo Batista
• Grantee: Renata Meyer Hobold
• Host Institution: Instituto de Matemática e Estatística (IME). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Company: Universidade Estadual de Campinas (UNICAMP). Faculdade de Engenharia Elétrica e de Computação (FEEC)
• Associated research grant: 21/00199-8 - SMART NEtworks and ServiceS for 2030 (SMARTNESS), AP.PCPE

Abstract

The constant evolution of new botnets oriented to cause Distributed Denial of Service (DDoS) attacks, combined with the low storage capacity of Internet of Things (IoT) devices to store packets traveling through the network, justify the use of intrusion detection systems (IDSs) based on machine learning models that are capable of learning about new attacks without the need to save all traffic. A good solution for this scenario is to use online learning, also known as stream learning. However, the low computational power of some IoT devices justifies that the model training should be performed in external environments with greater processing capacity, for example, through federated learning. Recently, some proposals in the literature have suggested using online federated learning. This scientific initiation project aims to evaluate the performance of different online federated learning models in detecting cyberattacks and to build a framework for experimenting with these models, allowing future research to be carried out without the need to write a new experimentation environment. The expected results are a report with the performance analysis of different online federated learning models and the framework code that will be made available as free and open-source software.