Next Article in Journal
Autonomous Greenhouse Cultivation of Dwarf Tomato: Performance Evaluation of Intelligent Algorithms for Multiple-Sensor Feedback
Next Article in Special Issue
WISEST: Weighted Interpolation for Synthetic Enhancement Using SMOTE with Thresholds
Previous Article in Journal
Time-Domain ADC and Security Co-Design for SiP-Based Wireless SAW Sensor Readers
Previous Article in Special Issue
A Hierarchical Dispatcher for Scheduling Multiple Deep Neural Networks (DNNs) on Edge Devices
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Efficient Keyset Design for Neural Networks Using Homomorphic Encryption †

1
Department of Electrical and Computer Engineering & ISRC, Seoul National University, Seoul 08826, Republic of Korea
2
Department of AI, Gachon University, Seongnam-si 13120, Republic of Korea
*
Authors to whom correspondence should be addressed.
This paper is an extended version of our paper published in “Rotation Keyset Generation Strategy for Efficient Neural Networks Using Homomorphic Encryption”. Presented at the International Conference on Artificial Intelligence Computing and Systems (AICompS) 2024, Jeju, Republic of Korea, 16–18 December 2024.
Sensors 2025, 25(14), 4320; https://doi.org/10.3390/s25144320
Submission received: 7 May 2025 / Revised: 3 July 2025 / Accepted: 8 July 2025 / Published: 10 July 2025
(This article belongs to the Special Issue Advances in Security of Mobile and Wireless Communications)

Abstract

With the advent of the Internet of Things (IoT), large volumes of sensitive data are produced from IoT devices, driving the adoption of Machine Learning as a Service (MLaaS) to overcome their limited computational resources. However, as privacy concerns in MLaaS grow, the demand for Privacy-Preserving Machine Learning (PPML) has increased. Fully Homomorphic Encryption (FHE) offers a promising solution by enabling computations on encrypted data without exposing the raw data. However, FHE-based neural network inference suffers from substantial overhead due to expensive primitive operations, such as ciphertext rotation and bootstrapping. While previous research has primarily focused on optimizing the efficiency of these computations, our work takes a different approach by concentrating on the rotation keyset design, a pre-generated data structure prepared before execution. We systematically explore three key design spaces (KDS) that influence rotation keyset design and propose an optimized keyset that reduces both computational overhead and memory consumption. To demonstrate the effectiveness of our new KDS design, we present two case studies that achieve up to 11.29× memory reduction and 1.67–2.55× speedup, highlighting the benefits of our optimized keyset.
Keywords: machine learning as a service; privacy-preserving machine learning; neural networks; fully homomorphic encryption; CKKS; rotation keyset; cryptography; privacy-preserving techniques; intrusion detection system machine learning as a service; privacy-preserving machine learning; neural networks; fully homomorphic encryption; CKKS; rotation keyset; cryptography; privacy-preserving techniques; intrusion detection system

Share and Cite

MDPI and ACS Style

Joo, Y.; Ha, S.; Oh, H.; Paek, Y. Efficient Keyset Design for Neural Networks Using Homomorphic Encryption. Sensors 2025, 25, 4320. https://doi.org/10.3390/s25144320

AMA Style

Joo Y, Ha S, Oh H, Paek Y. Efficient Keyset Design for Neural Networks Using Homomorphic Encryption. Sensors. 2025; 25(14):4320. https://doi.org/10.3390/s25144320

Chicago/Turabian Style

Joo, Youyeon, Seungjin Ha, Hyunyoung Oh, and Yunheung Paek. 2025. "Efficient Keyset Design for Neural Networks Using Homomorphic Encryption" Sensors 25, no. 14: 4320. https://doi.org/10.3390/s25144320

APA Style

Joo, Y., Ha, S., Oh, H., & Paek, Y. (2025). Efficient Keyset Design for Neural Networks Using Homomorphic Encryption. Sensors, 25(14), 4320. https://doi.org/10.3390/s25144320

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop