ANN Activation Function Estimators for Homomorphic Encrypted Inference

Homomorphic Encryption (HE) enables secure computations on encrypted data, facilitating machine learning inference in sensitive environments such as healthcare and finance. However, efficiently handling non-linear activation functions, specifically Sigmoid and Tanh, remains a significant computation...

Full description

Saved in:
Bibliographic Details
Main Authors: Mhd Raja Abou Harb, Baris Celiktas
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11034980/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850114120543633408
author Mhd Raja Abou Harb
Baris Celiktas
author_facet Mhd Raja Abou Harb
Baris Celiktas
author_sort Mhd Raja Abou Harb
collection DOAJ
description Homomorphic Encryption (HE) enables secure computations on encrypted data, facilitating machine learning inference in sensitive environments such as healthcare and finance. However, efficiently handling non-linear activation functions, specifically Sigmoid and Tanh, remains a significant computational challenge for encrypted inference using Artificial Neural Networks (ANNs). This study introduces a lightweight, ANN-based estimator designed to accurately approximate activation functions under homomorphic encryption. Unlike traditional polynomial and piecewise linear approximations, the proposed ANN estimators achieve superior accuracy with lower computational overhead associated with bootstrapping or high-degree polynomial techniques. These estimators are trained on plaintext data and seamlessly integrated into encrypted inference pipelines, significantly outperforming conventional methods. Experimental evaluations demonstrate notable improvements, with ANN estimators enhancing accuracy by approximately 2% for Sigmoid and up to 73% for Tanh functions, improving F1-scores by approximately 2% for Sigmoid and up to 88% for Tanh, and markedly reducing Mean Square Error (MSE) by up to 96% compared to polynomial approximations. The ANN estimator achieves an accuracy of 97.70% and an AUC of 0.9997 when integrated into a CNN architecture on the MNIST dataset, and an accuracy of 85.25% with an AUC of 0.9459 on the UCI Heart Disease dataset during ciphertext inference. These results underscore the estimator’s practical effectiveness and computational feasibility, making it suitable for secure and efficient ANN inference in encrypted environments.
format Article
id doaj-art-5f0b2b246de14156bbc45738b19b81eb
institution OA Journals
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-5f0b2b246de14156bbc45738b19b81eb2025-08-20T02:36:59ZengIEEEIEEE Access2169-35362025-01-011310351210353010.1109/ACCESS.2025.357966711034980ANN Activation Function Estimators for Homomorphic Encrypted InferenceMhd Raja Abou Harb0https://orcid.org/0000-0002-8637-5473Baris Celiktas1https://orcid.org/0000-0003-2865-6370Computer Engineering Department, Isik University, Istanbul, TürkiyeComputer Engineering Department, Isik University, Istanbul, TürkiyeHomomorphic Encryption (HE) enables secure computations on encrypted data, facilitating machine learning inference in sensitive environments such as healthcare and finance. However, efficiently handling non-linear activation functions, specifically Sigmoid and Tanh, remains a significant computational challenge for encrypted inference using Artificial Neural Networks (ANNs). This study introduces a lightweight, ANN-based estimator designed to accurately approximate activation functions under homomorphic encryption. Unlike traditional polynomial and piecewise linear approximations, the proposed ANN estimators achieve superior accuracy with lower computational overhead associated with bootstrapping or high-degree polynomial techniques. These estimators are trained on plaintext data and seamlessly integrated into encrypted inference pipelines, significantly outperforming conventional methods. Experimental evaluations demonstrate notable improvements, with ANN estimators enhancing accuracy by approximately 2% for Sigmoid and up to 73% for Tanh functions, improving F1-scores by approximately 2% for Sigmoid and up to 88% for Tanh, and markedly reducing Mean Square Error (MSE) by up to 96% compared to polynomial approximations. The ANN estimator achieves an accuracy of 97.70% and an AUC of 0.9997 when integrated into a CNN architecture on the MNIST dataset, and an accuracy of 85.25% with an AUC of 0.9459 on the UCI Heart Disease dataset during ciphertext inference. These results underscore the estimator’s practical effectiveness and computational feasibility, making it suitable for secure and efficient ANN inference in encrypted environments.https://ieeexplore.ieee.org/document/11034980/Activation function estimatorartificial neural networkencrypted inferencehomomorphic encryption
spellingShingle Mhd Raja Abou Harb
Baris Celiktas
ANN Activation Function Estimators for Homomorphic Encrypted Inference
IEEE Access
Activation function estimator
artificial neural network
encrypted inference
homomorphic encryption
title ANN Activation Function Estimators for Homomorphic Encrypted Inference
title_full ANN Activation Function Estimators for Homomorphic Encrypted Inference
title_fullStr ANN Activation Function Estimators for Homomorphic Encrypted Inference
title_full_unstemmed ANN Activation Function Estimators for Homomorphic Encrypted Inference
title_short ANN Activation Function Estimators for Homomorphic Encrypted Inference
title_sort ann activation function estimators for homomorphic encrypted inference
topic Activation function estimator
artificial neural network
encrypted inference
homomorphic encryption
url https://ieeexplore.ieee.org/document/11034980/
work_keys_str_mv AT mhdrajaabouharb annactivationfunctionestimatorsforhomomorphicencryptedinference
AT barisceliktas annactivationfunctionestimatorsforhomomorphicencryptedinference