Jonas Kuntzer

I am pursuing my Ph.D. at the Institute of Artificial Intelligence in Medicine at the Technical University of Munich (TUM). Before starting my doctoral studies, I worked as an applied scientist, where I focused on bridging the gap between research and industry in computer vision and NLP. Both my master’s and bachelor’s degrees were earned at the RWTH Aachen. For my master’s thesis, I specialized in enhancing audio signals by recovering them from noisy and filtered input.

My current research interests lie at the intersection of interpretability, privacy, and federated learning. Specifically, I investigate the interpretability of differentially private models, examining how their behavior diverges from non-private counterparts. In the realm of privacy, I explore the application of differential privacy and other privacy paradigms, such as data minimization and contextual integrity, focusing on handling unstructured data, including natural language and using large language models (LLMs). Furthermore, I am developing frameworks for federated learning to enable the secure and private extraction of knowledge from vast medical data archives across hospitals, strongly emphasizing integrating differential privacy to ensure robust privacy protections.
 

Interests:

• Interpretability
• Differential Privacy
• Federated Learning

Education:

• Electrical Engineering and Information Technology (M.Sc.), 2020
  RWTH Aachen
• Electrical Engineering and Information Technology (B.Sc.), 2017
  RWTH Aachen