On October 9 and 10, 2023, the Berlin Institute for the Foundations of Learning and Data (BIFOLD) at TU Berlin invited scientists from the university AI competence centers (BIFOLD, ScaDS.AI Dresden/Leipzig, Lamarr Institute, Tübingen AI Center and MCML) as well as the DFKI to Berlin. There they presented and discussed current results of their research. An AI symposium on the second day of the event presented the centers’ leading AI research to an interested public.
As part of the German government’s AI strategy, the five university-based AI competence centers have been receiving permanent funding from the federal government and their respective host state since July 1, 2022. Together with DFKI, they form a national network and are the basis of the German AI ecosystem. Parliamentary State Secretary to the Federal Minister of Education and Research Mario Brandenburg, who visited the centers and the DFKI this summer, said: “Each center has its own strengths. We want to build on this and develop them into leading European institutions for AI research.”
The AI competence centers form the foundation of the German AI ecosystem
October 9 was all about scientific exchange and networking. In addition to presentations from the centers, researchers came together in networking sessions and spoke on current topics in AI research – for example, trustworthy AI, Machine Learning that protects user privacy, and generative AI models. The latter were discussed intensively in a session led by Lamarr scientist Dr. Mehdi Ali. An AI symposium on October 10 was explicitly aimed at an interested public and brought together computer scientists, ethicists and politicians with experts from industry and representatives of society as well as over 200 guests. Researchers from the centers presented their specific research projects during a poster session.
NRW State Secretary Gonca Türkeli-Dehnert, who gave a welcoming speech on behalf of the funding institutions, said: “With the AI competence centers, Germany is excellently positioned to continue researching trustworthy AI for the benefit of humanity – and to do so at a top international level. By jointly funding the centers on a permanent basis, the federal and state governments are strengthening the link between research, teaching and transfer in the field of Artificial Intelligence and Machine Learning. This year’s AI Symposium in Berlin impressively demonstrated that these agile and innovative collaborations have an impact on society. ‘AI research made in Germany’ is a seal of quality for technology development worldwide.”
Sustainable positioning of the German AI landscape
What technical requirements must AI meet in order to be trustworthy? How can AI research be successfully transferred to industry? These and other questions were addressed and discussed by Principal Investigators of the Lamarr Institute in various panels. For example, Prof. Dr. Emmanuel Müller underlined the need for basic research on “Calibrated Trust”, where Machine Learning is equally perceived as trustworthy by humans but also ensures the necessary technical reliability.
Dr. Joachim Köhler highlighted the Lamarr Institute’s current research on foundation models, which have been in the focus of the public debate on AI not only since ChatGPT. Lamarr Director Prof. Dr. Christian Bauckhage emphasized in this context: “Scientists at the Lamarr Institute are working at full speed on generative AI as a prerequisite for the international competitiveness of Germany’s economy. We are proud to have already successfully developed large AI models – for example within the framework of the OpenGPT-X consortium project – and to have brought them into application.”
The event in pictures:
Copyright: Michael Setzpfandt
