Related Experiment Video
Updated: Jun 14, 2025
03:14
Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
Published on: December 6, 2024
523
Roadmap of Adversarial Machine Learning in Internet of Things-Enabled Security Systems
Yasmine Harbi1, Khedidja Medani1,2, Chirihane Gherbi1
1LRSD Laboratory, Ferhat Abbas University Setif-1, Setif 19000, Algeria.
Sensors (Basel, Switzerland)
|August 29, 2024
Summary
Adversarial attacks threaten machine learning (ML) in the Internet of Things (IoT). This study reviews adversarial machine learning (AML) in IoT security, highlighting the Fast Gradient Signed Method (FGSM) and recommending adversarial training for defense.
Area of Science:
- Cybersecurity
- Artificial Intelligence
- Internet of Things
Background:
- Machine learning (ML) and the Internet of Things (IoT) are pivotal in modern digital systems.
- The integration of ML and IoT enhances intelligent systems, particularly in security applications.
- ML-based IoT systems face significant vulnerabilities to adversarial attacks during model training and testing.
Purpose of the Study:
- To examine the severity and impact of adversarial attacks on ML-based IoT systems.
- To provide a comprehensive classification of adversarial machine learning (AML).
- To present a systematic literature review on AML and IoT security from 2020-2024.
Main Methods:
- Classification of adversarial machine learning techniques.
- Systematic literature review of recent research (2020-2024) on AML in IoT security.
- Analysis of common attack methods and defense strategies.
Main Results:
- Adversarial attacks pose substantial risks, including device malfunction and data misuse.
- The Fast Gradient Signed Method (FGSM) is identified as the most prevalent AML attack technique.
- Adversarial training is frequently recommended as a defense mechanism against these attacks.
Conclusions:
- Secure and robust ML models are crucial for IoT systems.
- Further research is needed to address open issues and enhance the security of ML-based IoT applications.
- Understanding AML threats is essential for developing resilient IoT security solutions.
