Making AI better together: The annual meeting of the SECAI Graduate School

From March 10 to 13, the third annual meeting of the SECAI “Zuse School of Excellence in AI” took place at the Schmochtitz educational estate. Almost 30 doctoral students from various research areas in artificial intelligence came together to exchange ideas and discuss current AI trends as well as scientific practice and career opportunities beyond the doctorate. “This meeting away from our offices is always a special highlight,” explains SECAI Director Prof. Markus Krötzsch (Faculty of Computer Science). “An intensive exchange takes place here - not only on a professional level.”

This time, a particular technical focus was on language-processing AI using so-called Large Language Models (LLMs), which are now known to a broad public thanks to ChatGPT & Co. In his guest lecture, Professor Simon Razniewski (Faculty of Computer Science and ScaDS.AI) discussed the question of what knowledge is contained in a language model and how - or whether - it can be distinguished from freely invented knowledge. Such research requires extensive, technically demanding experiments and Dr. Robert Haase (ScaDS.AI, Leipzig University) familiarized the researchers with the use of the LLM infrastructure in Dresden and Leipzig in an intensive course.

Another highlight was the introduction to the topic of founding a start-up as an attractive career path after the doctorate. Florian Sägebrecht presented launchhub42, the start-up incubator of TU Dresden and TU Munich, and Prof. Frank Fitzek and Justus Rischke (Soron Systems) shared their extensive experience as start-up founders. SECAI doctoral student Christian Vielhaus is also a co-founder of Soron Systems and thus perhaps also a pioneer for other researchers who venture the transfer into practice in the future.

The program was rounded off by Katharina Ulbrich (TUD Graduate Academy), whose workshop on good scientific practice conveyed the most important aspects of this complex topic and stimulated interesting discussions.

The School of Embedded Composite Artificial Intelligence (SECAI) is one of three DAAD Konrad Zuse Schools of Excellence in AI in Germany. The project of TU Dresden and Leipzig University promotes artificial intelligence (AI) in research and teaching, awards scholarships, strengthens education, funds researchers and promotes international exchange. Calls for applications for new doctoral positions will be published at https://secai.org/veröffentlichtat the beginning of May .