CEBIT 2018: Fraunhofer-Gesellschaft Presents AI Study

Machine Learning: Skills, Research, Application

Press release / June 07, 2018

Machine learning (ML) is the key technology for cognitive systems based on artificial intelligence (AI) and thus one of the decisive factors for global economic development. Fundamental for a sustainable positioning of Germany and Europe in international competition is the fact-based discussion of AI- and ML-based technologies. The new study of the Fraunhofer-Gesellschaft classifies the essential terms of machine learning, gives an overview of current challenges and future research tasks and presents Germany's position in the application of machine learning. The study will be officially presented at CEBIT 2018 (Hall 27, Booth E78).

© Fraunhofer IAIS / Fraunhofer IMW
Competence map based on scientific publications, 2006 - 2015.
© Fraunhofer IAIS / Fraunhofer IMW
Development of patent families on ML technology by country 2006 - 2015.

There is hardly an area that is not being decisively transformed by ML- and AI-based technologies: from goods production to logistics to medical technology. The sheer number of possible applications is one reason for the public interest. However, the debate is often characterized by half-knowledge, assumptions and myths. Clarification is needed, because social acceptance is of central importance for the further spread of machine-based learning methods. This is where the study "Machine Learning - Competencies, Applications and Research Needs", which was conducted in the context of a project funded by the German Federal Ministry of Education and Research (BMBF), comes in. The project was carried out by the Fraunhofer Institute for Intelligent Analysis and Information Systems IAIS, the Fraunhofer Center for International Management and Knowledge Economy IMW, and the headquarters of the Fraunhofer-Gesellschaft. The study provides a compact introduction to the most important concepts and methods of machine learning, an overview of challenges and new research questions. Furthermore, it provides an overview of actors, application fields and socio-economic framework conditions of research with a focus on Germany as a location.

Areas of Operation for Research

The scientific baseline situation in Germany and Europe is promising. However, the study identifies research areas that need to be intensified, particularly with regard to transfer into practice. The experts surveyed consider the following research fields to be particularly relevant, in which Germany and Europe should invest more in order to position themselves successfully in international competition in the long term.

  • "Explainable AI" for better transparency and reliability of ML-based decision processes
  • Machine learning with limited data
  • "Informed ML" - machine learning with additional knowledge from experts
  • Improving operational, cyber, and data security and robustness of
    ML-systems

These research fields offer the potential to expand knowledge in an application-oriented manner, enable completely new applications - from Industry 4.0 to the healthcare sector - and strengthen economic and social acceptance.

Key Challenges and Framework Conditions


Beyond the need for research in the central fields of action, overarching legal, social and political framework conditions also play an important role when it comes to the competitiveness and acceptance of ML and AI. The study particularly emphasizes the shortage of skilled workers, as the need for experts in data analysis in Germany is immense: There is currently a shortage of around 85,000 academics with advanced data analysis skills, as well as an additional 10,000 IT specialists in the fields of big data, advanced analytics, business intelligence and data science.


There is also a need for action with regard to the availability of data. Especially in international comparison, Germany lacks generally accessible, usable data. However, in order to create incentives to generate and share appropriate data, it is important that originators retain control and sovereignty over their data. Models such as the International Data Space, in which companies share their data for mutual benefit while always retaining control over the use of their data, are exemplary here.

Strategic Investment in ML Research


The Fraunhofer-Gesellschaft is not only committed to education and training with its own training and certification program for "Data Scientists ". At many institutes, both in Germany and in the European context, Fraunhofer develops key AI technologies and their applications. Machine learning methods for industry are just as much a part of this as the use of cognitive systems in cyber security and further research into artificial neural networks. In addition, Fraunhofer is addressing current industrial challenges, for example with the new Machine Learning Research Center, which enables transparent and resilient AI solutions to be strategically integrated into production, business and sales processes.


The study "Machine Learning - Competencies, Applications and Research Needs" will be presented at CEBIT 2018 (Fraunhofer booth Hall 27, booth E78). It is available for download (in German) in advance at: www.bigdata.fraunhofer.de/ml-studie