SecFedNIDS: Robust defense for poisoning attack against federated learning-based network intrusion detection system

Article Computer Science, Information Systems

D2MIF: A Malicious Model Detection Mechanism for Federated-Learning-Empowered Artificial Intelligence of Things

Wenxin Liu et al.

Summary: Artificial Intelligence of Things (AIoT), a fusion of AI and IoT, is a new trend in realizing industry 4.0 intelligentization, with data privacy and security being essential. In AIoT, federated learning is introduced to enhance data privacy protection by jointly training AI models without sharing private data. However, malicious participants can launch poisoning attacks, jeopardizing model convergence and accuracy. To solve this problem, we propose D2MIF, a malicious model detection mechanism based on iforest, which filters out models with malicious scores higher than a dynamically adjusted threshold using reinforcement learning. Experimental results on Mnist and Fashion_Mnist datasets show that D2MIF effectively detects malicious models and improves global model accuracy in federated learning-empowered AIoT.

IEEE INTERNET OF THINGS JOURNAL (2023)

Add to Collection

Article Computer Science, Artificial Intelligence

A scalable network intrusion detection system towards detecting, discovering, and learning unknown attacks

Zhao Zhang et al.

Summary: This study proposes a scalable intrusion detection system based on deep learning to detect, discover, and learn unknown attacks. The system includes an open-set classification network for detecting unknown attacks, a semantic embedding clustering method for discovering hidden unknown attacks, and an incremental nearest cluster centroid method for learning the discovered unknown attacks. Extensive experiments show that the system outperforms state-of-the-art methods in detecting various types of unknown attacks, demonstrating the feasibility of the proposed methods.

INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS (2021)

Add to Collection

Article Computer Science, Information Systems

Federated Learning With Unreliable Clients: Performance Analysis and Mechanism Design

Chuan Ma et al.

Summary: This article discusses the security risks that federated learning may encounter when training machine learning models among distributed clients, proposes a defensive mechanism named DeepSA, and validates its effectiveness through theoretical analysis and experimental results, comparing with other state-of-the-art defensive mechanisms.

IEEE INTERNET OF THINGS JOURNAL (2021)

Add to Collection

Article Automation & Control Systems

DeepFed: Federated Deep Learning for Intrusion Detection in Industrial Cyber-Physical Systems

Beibei Li et al.

Summary: The study introduces a novel federated deep learning scheme named DeepFed for detecting cyber threats against industrial CPSs. By designing a new intrusion detection model and federated learning framework, the research successfully achieves secure detection of various cyber threats.

IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS (2021)

Add to Collection

Article Computer Science, Theory & Methods

De-Pois: An Attack-Agnostic Defense against Data Poisoning Attacks

Jian Chen et al.

Summary: Machine learning techniques are vulnerable to data poisoning attacks, with current defense techniques largely attack-specific. De-Pois is proposed as an attack-agnostic defense, utilizing a mimic model trained to imitate the behavior of a target model trained by clean samples. By leveraging Generative Adversarial Networks (GANs), De-Pois is able to effectively detect poisoned data without explicit knowledge of ML algorithms or types of poisoning attacks.

IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY (2021)

Add to Collection

Article Computer Science, Information Systems