How important is German AI research worldwide?
Katharina Morik: Germany's research landscape is also well positioned by international standards: According to SCIMAGOjr1, Germany was in 6th place in the country comparison in terms of the number of AI publications both on average from 1996 to 2018 and in 2018 alone. In terms of the number of citations, Germany was even in 3rd place after the USA and UK. Thanks to the sustained support of the German Research Institute for Artificial Intelligence (DFKI) and several Max Planck and Fraunhofer Institutes, German research was able to assume a leading position early on. DFG Collaborative Research Centres also offer AI scientists a longer-term perspective and thus lead to outstanding research. Start-ups like DeepL or RapidMiner show how short the path from research to successful practice can be. By the way, the training of young researchers in Germany is so good that our postdocs are very popular all over the world.
What are the characteristics of Franco-German research cooperation?
Katharina Morik: There is a tradition of cooperation in the Franco-German border region, e.g. the Franco-German University in Saarbrücken or the agreement between DFKI and the French national institute for information technology (INRIA). Since Germany has not supported research on Machine Learning for a very long time, close cooperation with the French centres was of utmost importance for the development of this area in Germany. Together with the Department of Computer Science (LRI) of the University of Paris-Sud I was able to carry out European projects and a summer school on Machine Learning and "participants' practical problems" and thus helped to build up the field of Machine Learning in Germany as early as the 1980s. In the meantime, we are also establishing close cooperation with France through the AI competence centres in both countries. A first meeting was held in 2019, and more are planned. The aim is to create a network of centres that will result in a living ecosystem. Activities include the exchange of scientists for intensive about 4 weeks of joint thinking, the discussion of the curricula in AI and Data Science, as well as joint efforts for the transfer into applications and support with suitable computing capacity
Why is European cooperation necessary?
Katharina Morik: Europe has the potential to develop people-centred AI research that combines basic research with solid transfer into practice. The European conference ECML PKDD, for example, as a grassroots movement, has developed into a forum for exchange between scientists - 800 of them participated in 2019. In order to survive alongside China and the USA, however, a large scale, coordinated and long-term investment in research on Artificial Intelligence must be made in European countries. In China and the USA, about 20 professorships at a university are designated for sub-areas of AI and each is equipped with postdocs and computer infrastructure. AI research as an achievement of individuals is no longer effective.
 https://www.scimagojr.com/countryrank.php?category=1702&area=1700 Further reading: the Stanford AI index: https://hai.stanford.edu/sites/g/files/sbiybj10986/f/ai_index_2019_report.pdf