About me
I am a Researcher and Machine Learning Engineer, passionate about Medical Data Science and Explainable AI.
Medical Data Science
Seizure Prediction, X-Ray Segmentation, 3D Bone Reconstruction
ML Algorithms
Representation Learning, Deep CCA, Time-Series Anomaly Detection
Radiation Protection
Expertise R7: Operation of Medical X-Ray Equipment For Research Purposes
Experience
Since 2021
Paderborn University
Ph.D. Student
I research the subject of correlation-based deep multi-view representation learning, especially unsupervised methods like DeepCCA and its extensions. At the same time, I focus on combatting the "black-box" property of neural networks by developing practices of Explainable AI in combination with complexity bounds. That enables development of forward-thinking algorithms tackling fundamental medical problems like seizure prediction and computer-assisted diagnosis in general.
2020 - 2023
metamorphosis GmbH
Machine Learning Engineer
I was responsible for creating large databases of simulated X-ray images and training neural networks with them. I take pride in having developed a complete framework for GPU-accelerated X-ray image simulation for different surgical scenarios. Further, I trained, evaluated, and deployed multiple segmentation networks and developed algorithms that use that information for 3D reconstruction, image registration, and classification.
Achievements
2023
Paderborn University
Paderborn University research award 2023
I am happy to share that the project I am working in has won the Paderborn University Research Award 2023. The project is titled „Improving the lives of people with epilepsy: Towards a low-cost and real-time seizure prediction“. Together with our collaborators from the Sports Medicine Institute at Paderborn University and the division of Epilepsy and Clinical Neurophysiology at Boston Children’s Hospital, Harvard Medical School, we are working on an innovative solution for a wearable that offers seizure predictions in real time. An effective real-time prediction system could save lives by improving treatment strategies and even preventing seizures in the future. I am incredibly thankful to be part of this amazing research project and to be given the chance to work with all these talented researches. I am very excited about the future of the project! For further information, see the link to the press release: https://www.uni-paderborn.de/en/news-item/127266
Publications
Patents
- Applied: Artificial-Intelligence-Based Registration Of X-Ray Images
A. Blau, A. Lamm, M. Kuschel
metamorphosis GmbH, EP4014912A1, 2020
Journals
- Developing a deep canonical correlation-based technique for seizure prediction
S. Vieluf, T. Hasija, M. Kuschel, C. Reinsberger, T. Loddenkemper
Expert Systems with Applications, 2023
Conferences
- Rademacher Complexity Regularization for Correlation-based Multiview Representation Learning
M. Kuschel, T. Hasija, T. Marrinan
Accepted at International Conference on Acoustics, Speech and Signal Processing, IEEE, 2024 - Geodesic-Based Relaxation For Deep Canonical Correlation Analysis
M. Kuschel, T. Marrinan, T. Hasija
Workshop On Machine Learning For Signal Processing, IEEE, 2023 - FLight: FPGA Acceleration of Lightweight DNN Model Inference in Industrial Analytics
H. Mohammadi, F. P. Jentzsch, M. Kuschel, R. Arshad, S. Rautmare, S. Manjunatha, M. Platzner, A. Boschmann, D. Schollbach
Joint European Conference on Machine Learning and Knowledge Discovery in Databases, Springer, 2021 - DeepWind: An accurate wind turbine condition monitoring framework via deep learning on embedded platforms
H. Mohammadi, R. Arshad, S. Rautmare, S. Manjunatha, M. Kuschel, F. P. Jentzsch, M. Platzner, A. Boschmann, D. Schollbach
Conference on Emerging Technologies and Factory Automation, IEEE, 2020
Abstracts
- Multimodel pre-ictal changes in the autonomic nervous system in response to focal-onset epileptic seizures
T. Hasija, S. Vieluf, M. Kuschel, M. Jackson, S. Dailey, C. Reinsberger, T. Loddenkemper
AI in Epilepsy, 2024