Publications

You can also find my articles on my Google Scholar profile.

Accelerating Privacy-Preserving Medical Record Linkage: A Three-Party MPC Approach

Şeyma Selcan Mağara, Noah Dietrich, Ali Burak Ünal, Mete Akgün

Published in Journal of Biomedical Informatics, 2025

A novel and efficient Privacy-Preserving Record Linkage method based on a secure 3-party (MPC)computaion framework

Paper | Project Repo

Secure and Efficient Logistic Regression With Secret-Sharing MPC and Differential Privacy

Şeyma Selcan Mağara, Ipek Motorcu, Emin Şahin Mektepli, Cem Baykara, Ali Burak Ünal, Mete Akgün

Published in IEEE Access, 2025

Privacy-preserving approach for logistic regression that integrates secret sharing-based multi-party computation (MPC) with differential privacy (DP).

Paper | Project Repo

Privacy-Preserving Machine Learning (PPML) Inference for Clinically Actionable Models

Balaban, B., Magara, S. S., Yilgor, C., Yucekul, A., Obeid, I., Pizones, J., Kleinstueck, F., Perez-Grueso, F. J. S., Pellisé, F., Alanay, A., Savas, E., Bağcı, Ç., Sezerman, O. U., & European Spine Study Group

Published in IEEE Access, 2025

This paper presents a privacy-preserving inference framework for clinically actionable machine learning models, using homomorphic encryption to enable secure and accurate predictions.

Paper

Privacy Preserving Data Imputation via Multi-Party Computation for Medical Applications

Julia Jentsch, Ali Burak Ünal, Şeyma Selcan Mağara, Mete Akgün

Published in IEEE International Conference on E-health Networking, Application & Services (HealthCom), 2024

We propose a privacy-preserving data imputation framework using secure multi-party computation (MPC) to address missing values in medical datasets, ensuring regulatory compliance and maintaining model accuracy.

Paper

ML with HE: Privacy Preserving Machine Learning Inferences for Genome Studies

Mağara, Ş. S., Yıldırım, C., Yaman, F., Dilekoğlu, B., Tutaş, F. R., Öztürk, E., Kaya, K., Taştan, Ö., & Savaş, E

Published in ACM CCS 2021 PPML Workshop, 2021

A secure multi-label tumor classification method using homomorphic encryption, whereby two different machine learning algorithms, SVM and XGBoost, are used to classify the encrypted genome data of different tumor types.

Paper