Publications

2026

1. Temporal-Scale Sensitivity in Time-Series Tokenization and Scale-Robust Token Estimation by Gated Sum

   Berken Utku Demirel and Christian Holz

   In The Fortieth Conference on Neural Information Processing Systems (NeurIPS), 2026

   PDF (TO APPEAR)
2. InvisiSpeech: Making Silent Speech Invisible through a Miniature Intraoral Sensor

   Yijing Jiang, Taiting Lu, Berken Utku Demirel, Zhipeng Li, Mahanth Gowda, Michael Wessely, and Christian Holz

   In Proceedings of the ACM Symposium on User Interface Software and Technology (UIST), 2026

   PDF (TO APPEAR)
3. EgoHRV: Continuous Heart Rate Variability Estimation from Egocentric Systems for Autonomic Response and Skill Assessment

   Berken Utku Demirel and Christian Holz

   In European Conference on Computer Vision (ECCV), 2026

   arXiv PDF Code
4. Smartwatch-based Detection of Loss of Pulse

   Berken Utku Demirel, Galip Utku Akay, Paul Streli, Hatice Ozturkmen, and Christian Holz

   PLOS ONE, 2026

   PDF Code
5. Continuous Heart Rate Variability Estimation From PPG via State-Space Modeling

   Berken Utku Demirel and Christian Holz

   IEEE Transactions on Biomedical Engineering (IEEE TBME), 2026

   PDF Code
6. Temporal Cardiovascular Dynamics for Improved PPG-Based Heart Rate Estimation

   Berken Utku Demirel and Christian Holz

   IEEE Journal of Biomedical and Health Informatics (IEEE JBHI), 2026

   arXiv
7. FW-NKF: Frequency-Weighted Neural Kalman Filters

   Adnan Harun Dogan, Berken Utku Demirel, and Christian Holz

   In IEEE International Conference on Robotics and Automation (ICRA), 2026

   arXiv PDF Code

2025

1. Learning Without Augmenting: Unsupervised Time Series Representation Learning via Frame Projections.

   Berken Utku Demirel and Christian Holz

   In Thirty-ninth Conference on Neural Information Processing Systems (NeurIPS), 2025

   arXiv PDF Code
2. Beyond Subjectivity: Continuous Cybersickness Detection Using EEG-based Multitaper Spectrum Estimation

   Berken Utku Demirel, Adnan Harun Dogan, Juliete Rossie, Max Möbus, and Christian Holz

   IEEE Transactions on Visualization and Computer Graphics (TVCG), 2025

   arXiv Code
3. Shifting the Paradigm: A Diffeomorphism Between Time Series Data Manifolds for Achieving Shift-Invariancy in Deep Learning

   Berken Utku Demirel and Christian Holz

   In The Thirteenth International Conference on Learning Representations (ICLR), 2025

   arXiv PDF Code
4. Nightbeat: Heart Rate Estimation From a Wrist-Worn Accelerometer During Sleep

   Max Moebus, Lars Hauptmann, Nicolas Kopp, Berken Utku Demirel, Bjoern Braun, and Christian Holz

   IEEE Journal of Biomedical and Health Informatics (IEEE JBHI), 2025

   PDF Code

2024

1. Unobtrusive Air Leakage Estimation for Earables with In-ear Microphones

   Berken Utku Demirel, Ting Dang, Khaldoon Al-Naimi, Fahim Kawsar, and Alessandro Montanari

   Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. (IMWUT), 2024

   PDF
2. WildPPG: A Real-World PPG Dataset of Long Continuous Recordings

   Manuel Meier, Berken Utku Demirel, and Christian Holz

   In Advances in Neural Information Processing Systems (NeurIPS), 2024

   arXiv PDF Code
3. Tri-Spectral PPG: Robust Reflective Photoplethysmography by Fusing Multiple Wavelengths for Cardiac Monitoring

   Manuel Meier, Berken Utku Demirel, and Christian Holz

   In IEEE International Conference on Body Sensor Networks (IEEE BSN), 2024

   arXiv
4. An Unsupervised Approach for Periodic Source Detection in Time Series

   Berken Utku Demirel and Christian Holz

   In Forty-first International Conference on Machine Learning (ICML), 2024

   arXiv PDF Code
5. MiBOT: A Head-Worn Robot that Modulates Cardiovascular Responses through Human-Like Soft Massage

   Alice Mylaeus, Stephanie Vogt, Berken Utku Demirel, Marcel Gort, Mirko Meboldt, Manuel Meier, and Christian Holz

   In IEEE International Conference on Robotics and Automation (IEEE ICRA), 2024

2023

1. Finding Order in Chaos: A Novel Data Augmentation Method for Time Series in Contrastive Learning

   Berken Utku Demirel and Christian Holz

   In Thirty-seventh Conference on Neural Information Processing Systems (NeurIPS), 2023

   arXiv PDF Code
2. BeliefPPG: Uncertainty-aware Heart Rate Estimation from PPG Signals via Belief Propagation

   Valentin Bieri, Paul Streli, Berken Utku Demirel, and Christian Holz

   In Proceedings of the Conference on Uncertainty in Artificial Intelligence (UAI), 2023

   arXiv
3. Data-driven Energy-efficient Adaptive Sampling Using Deep Reinforcement Learning

   Berken Utku Demirel, Luke Chen, and Mohammad Abdullah Al Faruque

   ACM Trans. Comput. Healthcare (ACM Health), 2023

   Code
4. Cancelling Intermodulation Distortions for Otoacoustic Emission Measurements with Earbuds

   Berken Utku Demirel, Khaldoon Al-Naimi, Fahim Kawsar, and Alessandro Montanari

   In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (Top 3%), 2023

   PDF

   received a recognition award, placing in the top 3% among all the papers at the ICASSP

2022

1. AHAR: Adaptive CNN for Energy-Efficient Human Activity Recognition in Low-Power Edge Devices

   Nafiul Rashid, Berken Utku Demirel, and Mohammad Abdullah Al Faruque

   IEEE Internet of Things Journal (IEEE IoT-J), 2022
2. Energy-Efficient Real-Time Heart Monitoring on Edge–Fog–Cloud Internet of Medical Things

   Berken Utku Demirel, Islam Abdelsalam Bayoumy, and Mohammad Abdullah Al Faruque

   IEEE Internet of Things Journal (IEEE IoT-J), 2022
3. Neural Contextual Bandits Based Dynamic Sensor Selection for Low-Power Body-Area Networks

   Berken Utku Demirel, Luke Chen, and Mohammad Al Faruque

   In Proceedings of the ACM/IEEE International Symposium on Low Power Electronics and Design (IEEE/ACM ISLPED), Boston, MA, USA, 2022

   Providing health monitoring devices with machine intelligence is important for enabling automatic mobile healthcare applications. However, this brings additional challenges due to the resource scarcity of these devices. This work introduces a neural contextual bandits based dynamic sensor selection methodology for high-performance and resource-efficient body-area networks to realize next generation mobile health monitoring devices. The methodology utilizes contextual bandits to select the most informative sensor combinations during runtime and ignore redundant data for decreasing transmission and computing power in a body area network (BAN). The proposed method has been validated using one of the most common health monitoring applications: cardiac activity monitoring. Solutions from our proposed method are compared against those from related works in terms of classification performance and energy while considering the communication energy consumption. Our final solutions could reach 78.8% AU-PRC on the PTB-XL ECG dataset for cardiac abnormality detection while decreasing the overall energy consumption and computational energy by 3.7 \texttimes and 4.3 \texttimes, respectively.

4. Automatic Detection of the Spike-and-Wave Discharges in Absence Epilepsy for Humans and Rats Using Deep Learning

   Oguzhan Baser, Melis Yavuz, Kutay Ugurlu, Filiz Onat, and Berken Utku Demirel

   Biomedical Signal Processing and Control (BSPC), 2022
5. Template Matching Based Early Exit CNN for Energy-efficient Myocardial Infarction Detection on Low-power Wearable Devices

   Nafiul Rashid, Berken Utku Demirel, Mohanad Odema, and Mohammad Abdullah Al Faruque

   Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. (IMWUT), 2022

2021

1. Single-Channel EEG Based Arousal Level Estimation Using Multitaper Spectrum Estimation at Low-Power Wearable Devices

   Berken Utku Demirel, Ivan Skelin, Haoxin Zhang, Jack J. Lin, and Mohammad Abdullah Al Faruque

   In Annual International Conference of the IEEE Engineering in Medicine and Biology Society (IEEE EMBC), 2021
2. Two Might Do: A Beat-by-Beat Classification of Cardiac Abnormalities Using Deep Learning with Domain-Specific Features

   Berken Utku Demirel, Adnan Harun Dogan, and Mohammad Abdullah Al Faruque

   In Computing in Cardiology (CinC), 2021

2020

1. Quality Assessment of ECG Signals Based on Support Vector Machines and Binary Decision Trees

   Berken Utku Demirel and Yesim Serinagaoglu Dogrusoz

   In IEEE Signal Processing and Communications Applications Conference (SIU), 2020