Download PDFOpen PDF in browser

Federated Learning for Decentralized Intrusion Detection Systems (IDS)

EasyChair Preprint 13984

21 pagesDate: July 15, 2024

Abstract

In this research study, we investigate the application of federated learning for decentralized intrusion detection systems (IDS). Traditional IDS rely on a centralized architecture where all the data is collected and analyzed in a single location. However, this approach has several limitations, including privacy concerns and scalability issues. Federated learning, on the other hand, allows the training of machine learning models on distributed data without the need for data sharing.

 

Our research aims to explore the feasibility and effectiveness of using federated learning in IDS. We propose a decentralized architecture where multiple IDS nodes collaborate to collectively train a global intrusion detection model. Each node trains its local model on its own data while periodically exchanging model updates with other nodes. The global model is then updated by aggregating the local models' parameters.

 

To evaluate the performance of our proposed approach, we conducted experiments using a real-world dataset of network traffic. The results show that our federated learning-based IDS achieves comparable detection accuracy to the traditional centralized IDS, while addressing the limitations of data privacy and scalability. Furthermore, our approach demonstrates the potential for better generalization and adaptability, as the global model is trained on diverse data from different IDS nodes.

 

Overall, the findings of this research support the adoption of federated learning for decentralized intrusion detection systems. By leveraging the power of collaborative learning on distributed data, federated learning offers a promising solution for enhancing the effectiveness and scalability of IDS, while preserving data privacy.

Keyphrases: Cybersecurity, Internet, Technology

BibTeX entry
BibTeX does not have the right entry for preprints. This is a hack for producing the correct reference:
@booklet{EasyChair:13984,
  author    = {Kaledio Potter and Favour Olaoye and Lucas Doris},
  title     = {Federated Learning for Decentralized Intrusion Detection Systems (IDS)},
  howpublished = {EasyChair Preprint 13984},
  year      = {EasyChair, 2024}}
Download PDFOpen PDF in browser