Stefan Kalabakov is a PhD candidate whose research focuses on federated learning in the healthcare domain. Currently, his research is mainly focused on using federated learning to improve the generalization ability of clinical risk prediction models for Major Adverse Cardiovascular Events done in collaboration with Mt. Sinai (USA), Ribeirão Preto Medical School (Brazil), and The Styrian Hospital Association (Austria). This work is done under the PreCare-ML project.
He holds a Bachelor’s in Computer Engineering from the Faculty of Electrical Engineering & Information Technologies (Skopje) and a Master’s in Information and Communication Technologies from the Jožef Stefan International Postgraduate School (Ljubljana). During this time Stefan mostly worked on problems from the field of Human Activity Recognition using Inertial Measurement Units.
FedEHR-Bench: A Comprehensive Benchmarking Framework for Distributed Machine Learning on Electronic Health Records
Supervised Theses
Bachelor
2024
Theo Bardey
Comparing Federated Learning Personalization Strategies for Length of Stay Prediction on the eICU Collaborative Research Database
2024
Uli Prantz
Evaluating Foundation Model Performance on EHR Data from the eICU Database
2024
Julian Werne
Evaluating the Impact of Risk Communication Strategies in ClearRisk, PEER, ASCVD on Comprehension and Behavioral Intentions: A User Study
Master
2024
Marco Schaarschmidt
Benchmarking Federated Learning Methods on EHR Data
2025
Jan Carlo Schmidt
A comparison of EHR Foundation Models Against Expert-Defined Features for Predicting Crohn’s-like Disease of the Pouch
publications
Kalabakov, S. et al. A Comparative Analysis of Federated Learning for Speech-Based Cognitive Decline Detection. in Proc. Interspeech 2024 2455–2459 (2024).1.
Konak, O. et al. SONAR, a nursing activity dataset with inertial sensors. Scientific data10, 727 (2023).
Kalabakov, S. et al. Federated learning for activity recognition: A system level perspective. IEEE Access11, 64442–64457 (2023).
Kalabakov, S., Švigelj, A. & Javornik, T. Smartphone proximity detection using WiFi and BLE fingerprinting. in 2022 International Balkan Conference on Communications and Networking (BalkanCom) 36–40 (IEEE, 2022).
Kalabakov, S. et al. What actually works for activity recognition in scenarios with significant domain shift: Lessons learned from the 2019 and 2020 Sussex-Huawei challenges. Sensors22, 3613 (2022).
Kalabakov, S., Gjoreski, M., Gjoreski, H. & Gams, M. Analysis of deep transfer learning using deepconvlstm for human activity recognition from wearable sensors. Informatica45, (2021).
Gjoreski, H. et al. Head-AR: Human Activity Recognition with Head-Mounted IMU Using Weighted Ensemble Learning. in Activity and Behavior Computing (eds. Ahad, M. A. R., Inoue, S., Roggen, D. & Fujinami, K.) vol. 204 153–167 (Springer Singapore, Singapore, 2021).
Kalabakov, S. et al. Tackling the SHL challenge 2020 with person-specific classifiers and semi-supervised learning. in Adjunct Proceedings of the 2020 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2020 ACM International Symposium on Wearable Computers 323–328 (ACM, Virtual Event Mexico, 2020). doi:10.1145/3410530.3414848.
Gjoreski, M., Kalabakov, S., Luštrek, M., Gams, M. & Gjoreski, H. Cross-dataset deep transfer learning for activity recognition. in Adjunct Proceedings of the 2019 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2019 ACM International Symposium on Wearable Computers 714–718 (ACM, London United Kingdom, 2019). doi:10.1145/3341162.3344865.
