ANN activation function estimators for homomorphic encrypted inference

Basit Öğe Kaydı
dc.authorid0009-0001-4214-8738
dc.authorid0000-0003-2865-6370
dc.contributor.authorHarb, Mhd Raja Abouen_US
dc.contributor.authorÇeliktaş, Barışen_US
dc.date.accessioned2025-09-12T12:23:57Z
dc.date.available2025-09-12T12:23:57Z
dc.date.issued2025-06-13
dc.departmentIşık Üniversitesi, Lisansüstü Eğitim Enstitüsü, Bilgisayar Mühendisliği Yüksek Lisans Programıen_US
dc.departmentIşık University, School of Graduate Studies, Master’s Program in Computer Engineeringen_US
dc.departmentIşık Üniversitesi, Mühendislik ve Doğa Bilimleri Fakültesi, Bilgisayar Mühendisliği Bölümüen_US
dc.departmentIşık University, Faculty of Engineering and Natural Sciences, Department of Computer Engineeringen_US
dc.description.abstractHomomorphic 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.en_US
dc.description.versionPublisher's Versionen_US
dc.identifier.citationHarb, M. R. A. & Çeliktaş, B. (2025). ANN activation function estimators for homomorphic encrypted inference. IEEE Access, 13, 103512-103530. doi:https://doi.org/10.1109/ACCESS.2025.3579667en_US
dc.identifier.doi10.1109/ACCESS.2025.3579667
dc.identifier.endpage103530
dc.identifier.issn2169-3536
dc.identifier.scopus2-s2.0-105008277004
dc.identifier.scopusqualityQ1
dc.identifier.startpage103512
dc.identifier.urihttps://hdl.handle.net/11729/6699
dc.identifier.urihttps://doi.org/10.1109/ACCESS.2025.3579667
dc.identifier.volume13
dc.identifier.wosWOS:001512606800011
dc.identifier.wosqualityQ2
dc.indekslendigikaynakScopusen_US
dc.indekslendigikaynakWeb of Scienceen_US
dc.indekslendigikaynakScience Citation Index Expanded (SCI-EXPANDED)en_US
dc.institutionauthorHarb, Mhd Raja Abouen_US
dc.institutionauthorÇeliktaş, Barışen_US
dc.institutionauthorid0009-0001-4214-8738
dc.institutionauthorid0000-0003-2865-6370
dc.language.isoenen_US
dc.peerreviewedYesen_US
dc.publicationstatusPublisheden_US
dc.publisherInstitute of Electrical and Electronics Engineers Inc.en_US
dc.relation.ispartofIEEE Accessen_US
dc.relation.publicationcategoryMakale - Uluslararası Hakemli Dergi - Öğrencien_US
dc.relation.publicationcategoryMakale - Uluslararası - Kurum Öğretim Elemanıen_US
dc.rightsinfo:eu-repo/semantics/openAccessen_US
dc.subjectActivation function estimatoren_US
dc.subjectArtificial neural networken_US
dc.subjectEncrypted inferenceen_US
dc.subjectHomomorphic encryptionen_US
dc.subjectCiphertexten_US
dc.subjectClustering algorithmsen_US
dc.subjectInference enginesen_US
dc.subjectLearning systemsen_US
dc.subjectMachine learningen_US
dc.subjectMean square erroren_US
dc.subjectPiecewise linear techniquesen_US
dc.subjectPolynomial approximationen_US
dc.subjectPrivacy-preserving techniquesen_US
dc.subjectActivation functionsen_US
dc.subjectEncrypted inferenceen_US
dc.subjectHo-momorphic encryptionsen_US
dc.subjectHomomorphic-encryptionsen_US
dc.subjectMachine-learningen_US
dc.subjectNeural-networksen_US
dc.subjectPrivacy preservingen_US
dc.subjectPrivacy-preserving machine learningen_US
dc.subjectSigmoidsen_US
dc.subjectChemical activationen_US
dc.subjectNeural networksen_US
dc.subjectAccuracyen_US
dc.subjectPolynomialsen_US
dc.subjectComputer architectureen_US
dc.subjectComputational efficiencyen_US
dc.subjectHomomorphic encryptionen_US
dc.subjectDistributed databasesen_US
dc.subjectComputational modelingen_US
dc.subjectArithmeticen_US
dc.titleANN activation function estimators for homomorphic encrypted inferenceen_US
dc.typeArticleen_US
dspace.entity.typePublicationen_US

Dosyalar

Orijinal paket
Listeleniyor 1 - 1 / 1
Küçük Resim
İsim:
ANN_Activation_Function_Estimators_for_Homomorphic_Encrypted_Inference.pdf
Boyut:
2.45 MB
Biçim:
Adobe Portable Document Format
İndir
Lisans paketi
Listeleniyor 1 - 1 / 1
Küçük Resim Yok
İsim:
license.txt
Boyut:
1.17 KB
Biçim:
Item-specific license agreed upon to submission
Açıklama:
İndir

Koleksiyon

Öğrenci Yayınları Makale Koleksiyonu
Lisansüstü Eğitim Enstitüsü Diğer Yayınlar Koleksiyonu
Makale Koleksiyonu | Bilgisayar Mühendisliği Bölümü
Scopus İndeksli Yayınlar Koleksiyonu
WoS İndeksli Yayınlar Koleksiyonu