Investigating Adversarial AI Models and TinyML for Collaborative Cybersecurity Applications

• Grant number: 25/02406-1
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: April 01, 2025
• End date: February 28, 2026
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal Investigator: Michele Nogueira Lima
• Grantee: Gabriel Machado Violante
• Host Institution: Instituto de Ciências Exatas (ICEx). Universidade Federal de Minas Gerais (UFMG). Ministério da Educação (Brasil). Belo Horizonte , SP, Brazil
• Associated research grant: 18/23098-0 - MENTORED: from modeling to experimentation - predicting and detecting DDoS and zero-day attacks, AP.TEM

Abstract

The detection of malicious activities in networks has evolved over time, enhancing cybersecurity and making traffic monitoring more efficient. However, this task still faces significant challenges, particularly due to the computational limitations of edge devices, which hinder the application of traditional machine learning models. Given this scenario, the development of optimized solutions capable of operating efficiently in hardware-constrained environments becomes essential.Thus, this research project proposes the study and implementation of adversarial AI models in TinyML for monitoring and detecting malicious behaviors in networks using edge devices. Additionally, the project will explore the use of adversarial samples generated by generative AI models, both to enhance the robustness of these systems against evasion attempts and to potentially introduce adversarial traffic into the network as a defense strategy. The goal is to contribute to the advancement of attack detection strategies within the MENTORED project, strengthening cybersecurity through the application of machine learning techniques in embedded systems.