Institute of Computer Science
Heidelberg University
Im Neuenheimer Feld 205
69120 Heidelberg
Germany
Room 1/312
Hello!
I am Jonas Gann and am a PhD Researcher in the Data Science Group at Heidelberg University. My research begins with a simple question: How can we build AI systems that we do not merely trust, but can also understand and verify? I explore neuro-symbolic approaches that combine the flexibility of large language models with the transparency of logic-based reasoning, particularly for medical and other high-stakes applications.
Sometimes, the most trustworthy answer an AI can give is: “I do not know yet-tell me more.”
This idea is at the heart of my work. Instead of hiding uncertainty behind a confident answer, AI systems should be able to trace their conclusions back to reliable sources, make their true reasoning accessible, and recognize when important information is missing.
I am always happy to connect! Feel free to get in touch to exchange ideas, discuss research, or explore opportunities for collaboration.