Skip to content

Secure Federated Learning With Fully Homomorphic Encryption for IoT Communications

Authors: Neveen Mohammad Hijazi; Moayad Aloqaily; Mohsen Guizani; Bassem Ouni; Fakhri Karray
Year: 2024
Venue: IEEE Internet of Things Journal

type: application collection: ppml status: auto

Links: DOI

Summary

Auto-extracted summary (unreviewed)

The paper proposes secure federated learning for IoT-enabled smart cities by combining FL with fully homomorphic encryption so that model parameters are encrypted before transmission over insecure channels. Four FL-based FHE approaches are presented and evaluated; they retain high accuracy, recall, precision and F-score while cutting communication overhead by 80.15%-89.98% versus the baseline, and one variant lowers latency by 70.38%.

Abstract

The emergence of the Internet of Things (IoT) has revolutionized people’s daily lives, providing superior quality services in cognitive cities, healthcare, and smart buildings. However, smart buildings use heterogeneous networks. The massive number of interconnected IoT devices increases the possibility of IoT attacks, emphasizing the necessity of secure and privacy-preserving solutions. Federated learning (FL) has recently emerged as a promising machine learning (ML) paradigm for IoT networks to address these concerns. In FL, multiple devices collaborate to learn a global model without sharing their raw data. However, FL still faces privacy and security concerns due to the transmission of sensitive data (i.e., model parameters) over insecure communication channels. These concerns can be addressed using fully homomorphic encryption (FHE), a powerful cryptographic technique that enables computations on encrypted data without requiring them to be decrypted first. In this study, we propose a secure FL approach in IoT-enabled smart cities that combines FHE and FL to provide secure data and maintain privacy in distributed environments. We present four different FL-based FHE approaches in which data are encrypted and transmitted over a secure medium. The proposed approaches achieved high accuracy, recall, precision, and F-scores, in addition to providing strong privacy and security safeguards. Furthermore, the proposed approaches effectively reduced communication overhead and latency compared to the baseline approach. These approaches yielded improvements ranging from 80.15% to 89.98% in minimizing communication overhead. Additionally, one of the approaches achieved a remarkable latency reduction of 70.38%. The implementation of these security models is nontrivial, and the code is publicly available at https://github.com/Artifitialleap- MBZUAI/Secure-Federated-Learning-with-Fully-Homomorphic- Encryption-for-IoT-Communications.
Source: pdf

Extracted details

Field Value Source Confidence
contribution Four FL-based FHE approaches for secure parameter exchange in IoT smart-city settings llm 0.60
contribution Evaluation showing high accuracy/recall/precision/F-score with strong privacy guarantees llm 0.60
contribution Reduced communication overhead and latency compared with the baseline llm 0.60
contribution Publicly released implementation llm 0.60
schemes FHE llm 0.60
tasks federated learning llm 0.60
tasks IoT security / attack detection in smart cities llm 0.60
threat model FL devices and aggregation server communicating over insecure channels llm 0.60
interaction non-interactive FHE llm 0.60
code url https://github.com/Artifitialleap-MBZUAI/Secure-Federated-Learning-with-Fully-Homomorphic- (URL found in the full text) regex 0.50
code url https://github.com/Artifitialleap-MBZUAI/Secure-Federated-Learning-with-Fully-Homomorphic-Encryption-for-IoT-Communications (URL found in the full text) regex 0.50
code url https://github.com/Artifitialleap-MBZUAI/Secure-Federated-Learning-with-Fully-Homomorphic-Encryption-for-IoT-Communications llm 0.60
key result communication overhead reduction: 80.15%-89.98% (four FL+FHE approaches vs baseline) llm 0.60
key result latency reduction: 70.38% (best of the four approaches vs baseline) llm 0.60

Keyword mentions

Counts of vocabulary terms in the full text (first page of occurrence in brackets).

  • scheme: MPC 5 [p.2], Leveled / somewhat HE 4 [p.2], Differential privacy 3 [p.2], Partially homomorphic encryption 3 [p.2]
  • model: CNN 3 [p.5], RNN / LSTM 2 [p.12], MLP 1 [p.3], Autoencoder 1 [p.12]
  • dataset: UCI repository 1 [p.12], COVID 1 [p.12]
  • hardware: AMD EPYC 1 [p.7]
  • technique: Federated learning 25 [p.1], Bootstrapping 3 [p.3]
  • task: IoT / edge 78 [p.1], Medical / healthcare 19 [p.1], Clustering 3 [p.3]
  • topic: Homomorphic encryption 53 [p.1], Benchmark 3 [p.2], Open source 2 [p.1]

Files

Path (under papers/) Collection Pages Size SHA-256 Note
references/ppml/federated-learning/hijazi-2024_secure-federated-learning-with-fhe-for-iot-communications.pdf ppml/federated-learning 12 3.8 MB 876b9d822e61 was Hijazi et al. - 2024 - Secure Federated Learning With Fully Homomorphic Encryption for IoT Communications.pdf

Database key: secure-federated-learning-with-fully-homomorphic-encryption-for-iot · created 2026-11-05 · updated 2026-11-05 · edit with python3 tools/paperdb.py set secure-federated-learning-with-fully-homomorphic-encryption-for-iot COLUMN VALUE