DFKI News 45 e

N Like many employees around the world, research scientists at DFKI left the buildings in mid-March to begin working remotely. Re- search is still innovative even from your own home and project work has continued in a focused manner. After all, knowledge workers focus on knowledge and work, not on offices and buildings. The realm of ideas is found in the person: the brain also thinks when the person is at home, and it goes without saying that no one misses the commute to work or the search for a parking spot. Shockingly, a pandemic was necessary before many companies de- cided to seize the opportunities of digitalization consistently across the board and to explain the beneﬁts of distributed work to every- one. We have also had to learn completely new things with this strictly virtual collaboration and, in the process, discovered new as- pects of the corporate culture. DFKI project results are always team results, and individual projects may have very different durations. Collaborative research teams need spaces for meetings. Although recently, these spaces could no longer be physically co-located, our cooperation has still been quite productive. We installed col- laboration software at all locations and for all DFKI research depart- ments. Despite the physical separation, emotional and collegiate closeness was maintained, and that greatly facilitated our joint re- search and development efforts. We were excited about presenting the results at the Hannover Messe 2020. But, of course, the cancel- lation was necessary, and we have already registered our partici- pation for next year. That sends a clear signal to Hannover Messe and also to the builders of our exhibit stand! Together with our partners, we are developing AI to help the Ger- man and European economy and society get through this crisis, and emerge on the other side with a sustainable economic recov- ery and even a new ecological perspective! The EU funding outlook promises solid perspectives for European AI research. In mid-March, the EU Commission announced that it would provide EUR 50 million in start-up funding. Europe's largest AI initiative CLAIRE is tasked to coordinate scientiﬁc collaboration. COVID-19 has shifted the media spotlight away from the impor- tance of AI for the economy and society. But actually, it was the focus of a video conference on April 8 with Chancellor Angela Merkel, Federal Minister Anja Karliczek, State Minister Dorothee Bär, DFKI, and the heads of the ﬁve other German AI competence cen- ters. AI is helping to identify future pandemics earlier, giving politi- cians and economics more time to respond to new situations. AI can support the health system, regionalize supply chains, help cit- izens, facilitate neighborhood assistance, and strengthen cohesion. Our contribution to innovation continues. Stay creative and healthy! Prof. Dr. Antonio Krüger, DFKI CEO News 1_2020 I © DFKI 3

CONTENTS DIGITAL TRANSFORMATION Next Level AI: Hybrid AI Systems – The Best of Both Worlds p. 6 ROBOTICS Robotics Innovation Center – Robotic Systems for Production, Logistics, and People p. 14 5G Key Technologies in the Digital Transformation – Minister President Malu Dreyer Discusses With 5G and AI Experts p. 16 AI IN MANUFACTURING AI for Resilient Production – Project “SPAICER” Kick-off Event p. 33 Artiﬁcial Intelligence in Networked Production – Project AIRPoRT p. 34 Hybrid Teams for INDUSTRIE 4.0 – Research Project Hybr-iT Completed p. 35 The RICAIP-Center – INDUSTRIE 4.0 and AI as an Engine for Innovation in the Czech and German Economies p. 36 XR Works – Help for Assembly Line Workers p. 37 Demonstrator Construction of the SmartFactoryKL Continues p. 38 4 News 1_2020 I © DFKI COVID-19 App-based Analysis and Communication in Times of COVID-19 p. 8 Simulating Limited Contact to Fight the Coronavirus Pandemic p. 10 k c o t S e b o d A : o t o h P COOPERATIONS & NETWORKS Responsible Development of AI – Global Forum on AI for Humanity in Paris p. 18 Accelerating AI Together – DFKI Welcomes NVIDIA as Newest Shareholder p. 19 Inria and DFKI Sign Memorandum of Understanding for Cooperation in AI p. 20 CLAIRE Receives Broad Mandate and Funding for Shaping “AI Made in Europe” p. 22 Sartorius and DFKI Launch Joint Research Laboratory p. 24 PROJECT NEWS Safer Walker Prevents Falls Thanks to Intelligent Posture Recognition p. 25 Intelligent Environmental Protection and Smart Services from DFKI Laboratory Niedersachsen p. 26 LiDARshared – From the Networked Bus to Safe, Autonomous Use in Shared Space p. 27 Making AI Processes Transparent – Explaining Artificial Intelligence p. 28 Software Innovation for the Digital Transformation – Project MOSIM in the ITEA3 Program p. 30 AI and Data Analytics for Safer and More Reliable Mobility p. 30 QURATOR 2020 – From Experimental Concept to Routine Business p. 31

RESEARCH IN VIRTUAL SPACES p. 3 DIGITAL LEARNING New Learning Platform “School to go” p. 10 AI Campus – The Learning Platform for Artificial Intelligence p. 11 Opportunities in Digital Education – Interview With Prof. Dr. Niels Pinkwart p. 12 k c o t S e b o d A : o t o h P NEWS & EVENTS AI Research for Practical Use – DFKI Establishes Branch Office at the University of Trier p. 17 Professor Wahlster and Four DFKI Alumni Honored as Influential Thinkers in German AI Research p. 32 Imprint Issue 45, April 2020, ISSN 2196-2251 Published by: German Research Center for Artificial Intelligence GmbH (DFKI) Editorial staff: Heike Leonhard, Christof Burgard, Reinhard Karger, Armindo Ribeiro, Stephan Lehberger Editorial address: Saarland Informatics Campus D3 2, Stuhlsatzenhausweg 3, D-66123 Saarbrücken E-Mail: news@dfki.de, Phone: +49 681 85775 5390 Photos: DFKI, unless otherwise noted Coverphoto: Wall design at DFKI Project Office Berlin Layout, Graphics: Christof Burgard Translation team: Glenn Peach, Armindo Ribeiro, Sylvia Krüger, Heike Leonhard Production: One Vision Design Responsible: Heike Leonhard, Corporate Communications Frequency of publication: Semi-annual News online: www.dfki.de/en/dfki-news TABLE OF CONTENTS 3 Research in Virtual Spaces 6 Next Level AI: Hybrid AI Systems – The Best of Both Worlds 8 App-based Analysis and Communication in Times of COVID-19 10 Simulating Limited Contact to Fight the Coronavirus Pandemic 10 New Learning Platform “School to go” 11 AI Campus – The Learning Platform for AI 12 Opportunities in Digital Education – Interview With Prof. Dr. Niels Pinkwart 14 Robotics Innovation Center – Robotic Systems for Production, Logistics, and People 16 5G and AI are Key Technologies in the Digital Transformation 17 DFKI Establishes Branch Office at the University of Trier 18 Global Forum on AI for Humanity in Paris 19 DFKI Welcomes NVIDIA as Newest Shareholder 20 Inria and DFKI Sign Memorandum of Understanding for Cooperation in AI 22 CLAIRE Receives Broad Mandate and Funding for Shaping “AI Made in Europe” 24 Sartorius and DFKI Launch Joint Research Laboratory 25 Safer Walker Prevents Falls Thanks to Intelligent Posture Recognition 26 Intelligent Environmental Protection and Smart Services from DFKI Laboratory Niedersachsen 27 LiDARshared – From the Networked Bus to Safe, Autonomous Use in Shared Space 28 Making AI Processes Transparent – Explaining Artificial Intelligence 30 Software Innovation for the Digital Transformation – Project MOSIM in the ITEA3 Program 30 AI and Data Analytics for Safer and More Reliable Mobility 31 QURATOR 2020 – From Experimental Concept to Routine Business 32 Professor Wahlster and Four DFKI Alumni Honored as Influential Thinkers in German AI Research 33 Project “SPAICER” Kick-off Event 34 Artificial Intelligence in Networked Production – Project AIRPoRT 35 Hybrid Teams for INDUSTRIE 4.0 – Research Project Hybr-iT Completed 36 The RICAIP-Center – AI for INDUSTRIE 4.0 37 XR-Works – Help for Assembly Line Workers 38 Demonstrator Construction of the SmartFactoryKL Continues 39 DFKI Interview – Simon Bergweiler 40 News in Brief 42 Company Profile News 1_2020 I © DFKI 5

Prof. Dr. Antonio Krüger. Technology companies and states are invesng billions in the research, development, and use of arﬁcial intelligence. Europe needs to deﬁne for itself what kind of AI systems will be used in ten, ﬁoeen, or twenty years' me and what the corresponding AI tools should achieve. Hybrid AI systems are the core of "human-centric AI," which is human-oriented in its manifestaon and has the potenal to become a worldwide export hit. By combining the advantages of machine learning with the possibility of explainability through the formal inclusion of domain mo- deling and expert knowledge, the increased economic, social, polical, and liability law requirements for an "AI made in Europe" can be met. N The further development and use of AI results in Germany re- quire that the stakeholders in science, research, and industry be- come networked also at the European level, so that the critical mass of talent, investment volume, and entrepreneurship can be established and maintained. At the start of 2020, DFKI employed more than 1,000 people. Yet, the necessary norms and standards in basic and applied research will only be developed and deﬁned via networks – not only research networks but also by taking indus- try and professional associations into account. The results of the application-oriented basic research must be applied and lead to added value in innovative products and services in Europe. European cross-border cooperation is required. The initial steps have been taken, for example, the CLAIRE and ELLIS initiatives, which connect all of the AI research centers in Europe and focuses on human-centric AI applications. A lighthouse project is clearly needed at the European level, as recently proposed by the EU Com- mission. Such an AI hub should not be designed as a central mega cam- pus and research jug- gernaut but as a large and recognized international meeting place for AI scientists and users, where science and industry can tem- porarily work together on future-oriented AI topics. The strength of decentralized AI research in Europe must also be further consoli- dated by creating regional beacons with thematic focuses and net- working these in terms of content. Europe should join forces and create large AI test centers and environments together with indus- try, where the latest research results can be tested under realistic conditions. European AI and DFKI have been covering the entire spectrum of artiﬁcial intelligence for decades. The quality of research in Ger- many has been and continues to be high, with the intellectual ex- panse given and a variety of methods practiced. However, it is true that machine learning, a subject of intense coverage lately in the media, plays a less prominent role in Germany than in Silicon Valley. Neural networks have been around since the 1960s, but because of the low processing power and insufficient data, this approach had little impact on industry. True, this use of AI in the service sector is dominated by the globally operating US Internet companies. But in the use of AI in production technologies and INDUSTRIE 4.0, 6 News 1_2020 I © DFKI

Germany is especially signiﬁcant, including the areas of au- tonomous driving or business software. In other words, areas where the decisive criteria for use are: quality of products, achievement of objectives, and reliability. Even as breakthroughs in machine learning dominated the head- lines and conferences over the past ten years, DFKI never did switch course completely. A monolithic approach cannot deliver the per- formance required for the increasing roles, tasks, and functional scope of AI systems. This can only be achieved by systems that combine the use of statistical methods such as classical machine learning, as well as sub-symbolic deep learning, with symbolic ap- proaches based on formal logic and models of the subject matter, which may come from domain or expert knowledge. Only these hybrid systems will be able to meet the increased economic, soci- etal, political, and legal liability requirements. The human expert can dock and add factual and empirical knowledge. Hybrid meth- ods are becoming more important for a variety of reasons. These reasons include the need to explain system results, predict future results, and track performance, which are necessary prerequisites for certiﬁcation of systems. Important projects at DFKI address the “explainability” of AI sys- tems and help to increase conﬁdence in AI tools. When a mail- order company, for example, recommends the “wrong” book, i.e., one that does not interest the customer, this is emotionally unsat- isfactory. When that happens, it's not too serious and doesn't in- volve ﬁnancial or health problems. The wrong diagnosis by a medicine algorithm, in turn, can have fatal consequences. Such outcomes must be questioned, and it must be possible to question these assessments in such a way that the doctor understands how the result came about. “The AI systems of the future will provide better support to people in more and more sensitive areas. But the ﬁnal decision will always be made by a human. Nevertheless, a one hundred percent safety level is not achievable, and a bit of realism and modesty is important. No AI system and no doctor will have an error rate of zero. This issue can only be managed through open discussion in a social dialog. The German Bundestag's investigative body for Artiﬁcial Intelligence (Enquete-Kommission), the government's Data Ethics Commission, and the European Commission's High-Level Expert Group on AI are all doing important work in this regard.” Prof. Dr. Antonio Krüger MORE INFORMATION Ħ www.dfki.de CONTACT Ü Prof. Dr. Antonio Krüger DFKI CEO, Technical and Scientiﬁc Managing Director Ö Antonio.Krueger@dfki.de 4 +49 681 85775 5006 k c o t S e b o d A : o t o h P News 1_2020 I © DFKI 7

App-based Analysis and Communication in Times of COVID-19 The eﬀorts to contain the spread of the coronavirus pandemic increasingly focus on the use of smartphone apps. Used correctly, such apps may have the potenal to provide valuable informaon on the spread of infecons and to contribute signiﬁcantly to reducing the rate of new infecons. Based on proven crowd sensing technologies, a team at DFKI, together with its spin-oﬀ company SIS Sooware, has designed a system that records the relevant health parameters and enables contact tracing based on the European PEPP-PT standard communicaon opons, while respecng privacy and data protecon laws. “Artiﬁcial intelligence can not only help to detect the outbreak and spread of epidemics at an early stage through pattern recognition, but can also contribute to efficient control measures. How to collect the underlying data and make it available for crisis management with due consideration of data protection has been the subject of groundbreaking EU research projects in the past and now proven in practical use.” Prof. Dr. Paul Lukowicz, Head of Research Department Embedded Intelligence The system comprises the following core components: Recording of health and analysis parameters The health parameters voluntarily provided by users are anonymized, aggregated, and analyzed to map the exact spatial distribution of infection. The parameters to be queried are deﬁned by leading Eu- ropean epidemiologists of the inﬂuenza monitoring network “ln- ﬂuenzaNet” and adapted over time to be stay relevant to the current situation of the users. These parameters are not associated with any personal data. This project beneﬁts from experience from the EU-funded research project CIMPLEX, which developed various tools to investigate and inﬂuence the spread of disease and other infection phenomena in complex social systems. THE MAIN FUNCTIONS OF THE SIS COVID-19 SYSTEM Respect privacy and allowing voluntariness Recording of health and analysis parameters Decentralized, anony- mous contact tracing Anonymized analysis of mobility patterns Geo-based dispatch of messages Marking of important infrastructure in the app Option for official requirements: permit High scalability regard- ing the number of users 8 News 1_2020 I © DFKI

In addition to DFKI spin-off SIS Software, the DFKI team includes the business consultancy PricewaterhouseCoopers (PwC), the concert promoter Eventim, and the start-up Coneno. The European inﬂuenza tracking network lnﬂuenzaNet, which organizes leading epidemiol- ogists, is also a member of the team. s Project information at a glance SIS Software GmbH was founded in 2012 as a spin-off of the German Research Center for Artiﬁcial Intelligence (DFKI). The company is the market leader in the ﬁeld of digital crowd management and is primarily concerned with the analysis and notiﬁcation of large crowds. The company's clientele ranges from security authorities such as the Dutch police, to major events such as “Rock am Ring”, to the business consul- tancy PwC, and the international automobile association FIA. Project Partners: • PricewaterhouseCoopers (PwC) • Eventim • European Inﬂuenza Tracking Network “lnﬂuenzaNet” • Coneno MORE INFORMATION Ħ www.dfki.de/ei Ħ www.sis-software.de CONTACT DFKI Ü Prof. Dr. Paul Lukowicz Head of Research Department Embedded Intelligence Ö Paul.Lukowicz@dfki.de 4 +49 631 20575 4000 CONTACT SIS SOFTWARE Ü Dr. Tobias Franke SIS Software GmbH Nördliche Münchner Straße 27A 82031 Grünwald Germany Ö info@sis-software.de Decentralized, anonymous contact tracing Using so-called “Bluetooth handshakes,” anonymous contact trac- ing is achieved without sending data to a server so as to protect the privacy of the users. This technology performs an important task in the interruption of infection chains by quickly warning users when they have been in contact with a positively diagnosed person. As an early adopter, the team is in the process of adapting the sys- tem to the European PEPP-PT contact tracing standard, which is currently under development. This EU standard provides a decen- tralized, anonymous, and logbook-based contact tracing guidance, in full compliance with the General Data Protection Regulation (GDPR). It also ensures international compliance so that chains of infection can be traced across national borders. Targeted communication Messages and notices can be sent to location-speciﬁc areas, so- called “location-based messages. These are only received in cer- tain geographical areas. The shape and size of the message target area can be selected as required: every scenario is possible – from single blocks of houses, to public places, even entire countries. In the context of a pandemic, the population can be provided with targeted information and tips – nationwide, locally, by region, or by event organizer. In this way, users can be speciﬁcally notiﬁed if they are at particular risk of infection. Optional anonymized analysis of mobility patterns The crowd sensing technology was developed within the frame- work of the EU research project SOCIONICAL, in which 14 European universities and research centers investigated the interplay of tech- nology and social interaction. It was initially developed and used to analyze the behavior of crowd movements at major events and to prevent catastrophes, such as the 2010 Love Parade in Duisburg. These analyses are based on anonymous, aggregated movement data sent with the consent of smartphone users. At major events such as the 2012 Olympic Games in London or the “Rock am Ring” Festival, the technology has already been success- fully used to prevent potential disasters. In a crisis, this technology can be used to analyze mobility patterns and draw conclusions about how the epidemic is spreading. “Our system aims to develop a holistic platform for citizen participation, data analysis and official information. We have integrated the results of successful research projects and proven technolo- gies. In the context of the current COVID-19 crisis, these can make an important contribution to combating the pandemic. The components can be used individually as well as integrated in combination with other features.” Dr. Tobias Franke, Scientist in the Embedded Intelligence Research Department DFKI-Kaiserslautern and founder of SIS Software News 1_2020 I © DFKI 9

Simulating Limited Contact to Fight the Coronavirus Pandemic N What happens when a school that has been closed because of COVID-19 is reopened? What is the impact if just under half of the population – mostly the younger people – uses an app designed to help contain the coronavirus? Scientists in the information man- agement departments of DKFI and Trier University are using Artiﬁ- cial Intelligence to create pandemic scenarios that answer these questions. These researchers enter various parameters, such as disease pro- gression and people's behavioral patterns into simulation models. These simulations calculate how a virus can spread as people meet at work, in schools, or during their free time. The results can help in assessing which measures are useful and which are not. Several years ago, Professor Ingo Timm, of the Information Man- agement department at Trier University and head of Cognitive So- cial Simulation research at DFKI, started working on simulation models for the spread of the ﬂu. He and a colleague, Dr. Jan Ole Berndt, are now applying AI methods to the example of COVID-19 to calculate various pandemic responses. “Computer simulation has great potential in the management of pandemics. As part of our research, we integrate knowledge and expertise from medicine, sociology, and psychology. The spread of the virus depends on the behavior of people and factors such as Besides pandemics, Prof. Ingo Timm also researches other public safety issues, such as the design of escape routes in the event of gun rampages. school closings. We simulate these in our models with the aim of improving our ability to assess the effectiveness of these measures,” said Timm. MORE INFORMATION Ħ www.dfki.de/sds CONTACT Ü Prof. Dr. Ingo Timm Head of Cognitive Social Simulation Research Ö Ingo.Timm@dfki.de 4 +49 651 201 2859 New Learning Platform “School to go” N The nationwide school closings have given pupils an unexpect- edly large amount of free time. DFKI, together with the Bildung Digital Research Institute and the Didactic Innovations company, hope that the “School to go” learning platform will help pupils to use this time as sensibly as possible and motivate them to try digital learning. “School to go” went online on March 2020, and bundles innovative learning opportunities for children, adolescents, and young adults. It is the ﬁrst platform for modern learning that works like a social media outlet. A 20-member team of didactics and technologists is currently working to provide motivating, curriculum-compliant content on this platform. As a non-proﬁt solution, quality-assured learning opportunities for every subject and age are available free of charge to pupils, parents, teachers, and the interested public. The courses are sorted according to target age groups, for example, 6- 10 years, 11-14 years, and 15-18 years. Subject areas include German, foreign languages, MINT, and a learning mix, and consist of inno- vative learning formats (e.g., interactive videos, micro-learning el- ements, multimedia presentations). The initiative is managed and implemented by Prof. Dr. Julia Knopf at the Research Institute for Digital Education (FoBiD), in collabo- ration with Didactic Innovations and DFKI's Smart Enterprise Engi- neering department headed by Prof. Dr. Oliver Thomas. This initiative is supported by the Didacta Association as a cooperation partner. MORE INFORMATION Ħ https://schooltogo.de CONTACT Ü Prof. Dr. Oliver Thomas Head of Research Department Smart Enterprise Engineering Ö Oliver.Thomas@dfki.de 4 +49 541 386050 4830 10 News 1_2020 I © DFKI

AI Campus – The Learning Platform for Artiﬁcial Intelligence “AI Campus – The learning pla|orm for Arﬁcial Intelligence” is a pilot project launched and sponsored by the Federal Ministry of Educaon and Research (BMBF) in October 2019. The aim of the 3-year pilot phase is to develop a prototype of a digital learning pla|orm that is spe- cialized in arﬁcial intelligence. The AI Campus is implemented by the Soerverband, DFKI, Hasso Planer Instute (HPI), NEOCOSMO, and the mmb Instute. N The learning platform is designed for students, the employed, and other lifelong learners who have an interest in artiﬁcial intelli- gence. The AI campus is also aimed at universities, research insti- tutions, and companies that want to participate with their own learning curricula or to integrate content from the AI campus into their learning curricula. The purpose is to impart AI skills and get more people excited about the subject. Students will be taught to understand, question, and participate in the advances being made in AI. The AI Campus will contribute to the alleviation of the acute shortage of skilled staff while, in a broader sense, strengthening a responsible ap- proach to artiﬁcial intelligence. To provide students with the best possible courses, classes are spe- cially developed in house for the AI Campus with some existing (open) learning classes also being integrated. DFKI produces several of its own courses for the AI Campus. Additional learning classes will be developed in the context of competitions and collabora- tions. An initial competition for the development of innovative learning ideas met with great interest: As of February 2020, universities, re- search institutes, and companies had submitted a total of 137 pro- posed ideas and applied for BMBF funding. Funding for the ﬁrst projects selected as winners in the competition begins in July 2020. Course content at the AI campus will include the fundamentals as well as interdisciplinary questions and in-depth coverage of indi- vidual topics such as deep learning. In addition, speciﬁc offers are to be created that target the needs of distinct professional groups and specialist areas, for example, AI in the medical ﬁeld. All course offerings and systems at the AI Campus follow the principle of open resources and source codes. DFKI is also developing a series of AI-based features for the AI Cam- pus to include intelligent and self-directed learning. A prototype of the AI Campus platform with the starting content is expected to be available in mid-2020 and will then be successively expanded to include additional learning courses and functionalities. MORE INFORMATION Ħ www.ki-campus.org CONTACT Ü Dr. Kinga Schumacher Research Department Cognitive Assistants Ö Kinga.Schumacher@dfki.de 4 +49 30 23895 1819 Graphic: KI-Campus News 1_2020 I © DFKI 11

Opportunities in Digital Education – Interview With Prof. Dr. Niels Pinkwart 9 What do you say to your neighbors when they ask you what you do for a living? Prof. Dr. Niels Pinkwart: I tell them that I am working on a very ex- citing interface – namely, one between digitalization and educa- tion. I am a researcher investigating the choices AI can offer in various contexts for certain learning and educational processes. 9 Digital education is not a new subject. What can AI contribute? Prof. Dr. Niels Pinkwart: Digitalization and education have a long history together in different, interconnected ways. For example, computer science education has long been focused on sharing knowledge about digital technologies. Knowledge of AI technolo- gies, which are affecting our lives in more and more different ways, is becoming more and more important to us all as human beings. We must all learn and understand what AI does to us, for us, and with us. AI is then seen as educational content. AI methods may also be viewed as educational media. The use of AI technologies in education gives us the chance to design learning scenarios much more efficiently, for example, by responding to the preferences and previous knowledge of an individual learner. Such scenarios demand completely new types of interfaces, which in- deed cover things like stress during learning and emotions. The re- sult is a learning experience and learning outcomes that are much more tangible. 9 About the teaching aspect of AI: DFKI is a partner in the AI Cam- pus, a pilot project sponsored by the Federal Ministry of Education and Research (BMBF). What is your role personally and that of the Educational Technology Lab (EdTec)? Prof. Dr. Niels Pinkwart: The AI Campus is a project that includes EdTec as well as other institutes and several DFKI research depart- ments. Shortly after I started working at DFKI Berlin, I was given re- sponsibility as the scientiﬁc director of the AI Campus project. The EdTec Lab has two functions at the AI Campus: Our lab is designing a course in AI and Leadership that will soon be produced and made available to the public. We are also researching the duality between educational content and medium and how AI can be used to add value to user interac- tions, for example, to improve the interactions with a chatbot. In the course of the chat, the user can express what the educational need is and receive a personalized course that consists of very dif- ferent elements. These courses can be hosted on the AI Campus platform and may include external educational resources. The EdTec Lab at DFKI develops AI-based software technologies for training, qualification, and educational processes in school, academic, and vocational education and training. Photo: Adobe Stock 12 News 2_2019 I © DFKI

In practice, we still have a long way to go. Yet, many teachers are now out of necessity forced to say: “We're going to ﬁnd out how to use digital tools to design lessons.” This is the case in schools and also in universities. We have been told in queries from Humboldt University that an estimated 90% of courses in this situation can be conducted digitally. The current momentum sweeping everyone along with it implies an enormous potential for change. The teach- ers who participated in the webinar found the insights to be inspir- ing and conﬁrmed that their use of the technologies would deﬁnitely continue. It will certainly be similar at universities. If one consequence of the corona crisis is a higher level of dissem- ination and acceptance of these technologies, the next steps will be personalization, automated recommendations for the next learning step, etc. Soon AI will offer additional functions that make life easier and learning even more pleasant and efficient. The hur- dles are now lower since the teachers at schools and universities will certainly not set aside something once they have had a positive experience with it. The EdTec Lab hopes to be involved, of course, in an advisory capacity and as a partner for educational institutions in future research and industry projects. MORE INFORMATION Ħ www.dfki.de/edtec CONTACT Ü Prof. Dr. Niels Pinkwart Head of Research Department Educational Technology Lab Ö edtec@dfki.de 4 +49 30 23895 0 News 1_2020 I © DFKI 13 Prof. Niels Pinkwart. Photo: Jan Kopankiewicz 9 EdTec Lab is also working on the KI.ASSIST project. Can you brieﬂy outline what that is all about? Prof. Dr. Niels Pinkwart: KI.ASSIST is another exciting project that covers a very important aspect: Inclusion. The more educational scenarios become digitized, the more important it is to actually in- volve all people with all their individual characteristics, including possible disabilities. Learning systems must be designed that ex- plicitly include everyone and, in particular, use the tools of AI to provide digital assistance services. Simple AI aided systems are already in use to support visually im- paired people, for example, through speech recognition and read- ing functions. Project KI.ASSIST is collecting established assistance systems for people with handicaps and working to develop appli- cation scenarios for vocational rehabilitation. We look at how people with and without disabilities can use AI- based systems. In the follow-on phase, we want to establish exper- imental learning spaces at partner facilities and workshops and look at how we can initiate transformative processes. 9 The coronavirus crisis has apparently triggered a boom in digital education. What is your assessment? Do you think it is sustainable? Prof. Dr. Niels Pinkwart: My team came up with the idea of setting up a webinar on short notice for Berlin's teachers, explaining the basics of online-teaching. Thanks to the Berlin Senate’s support, we had more than 2,000 registrations in two days. Ninety-ﬁve percent of the participants stated that they had never, or very rarely, worked with digital media in school before. The demand is great, as many schools initially had to deal with the emergency with ready solu- tions, such as scanning worksheets and sending them to the pupils by e-mail. To a certain extent, this runs counter to the AI-assisted teaching models where each student receives personalized feed- back on their assignments.

The mobile manipulator “Mobipick” hands a cordless screwdriver to an Airbus employee as part of the Hybr-iT project for improved human-robot collaboration. Robotics Innovation Center – Robotic Systems for Production, Logistics, and People A world without industrial robots is hard to imagine today. In Germany alone, the use of robots in the manufacturing sector increased by 26 percent in 2018. These systems are being used more and more, mainly, in the automove and electronics industries, but also in other areas of producon and logiscs. At the German Research Center for Arﬁcial Intelligence, people play a major role in this development: the current research on robocs is not placing people on the outside, but rather at the center of desirable innovaon. N In this context of “Human-Centric AI,” the projects at the Robotics Innovation Center in Bremen are about supporting the employee. The industrial use of robots and artiﬁcial intelligence in the future poses questions as numerous as the po- tential beneﬁts for people. The core of the research effort is on the cooperation between people and machines, which, when implemented correctly, avoids health risks and makes dangerous work steps a thing of the past. In addition to robot systems for underwater, space, and rehabilitation use, the Bremen research division of DFKI is also developing systems that can support hu- mans and act autonomously in industrial production environments. Intelligent human-robot collaboration, which allows work in hybrid teams without protective wall dividers. This inherently requires safe robotic systems that enable reliable in- teraction with humans, as well as intuitively programmable systems that can be used in ﬂexible low-volume production. Working together with other DFKI research departments and industrial partners such as Airbus and Volkswagen, the Robotics Innovation Center is researching this form of hybrid and intelligent human-robot collaboration for the application areas of vehicle and aircraft assembly in project Hybr-iT. An important aspect of this effort involves worker recognition in real-time, which is necessary for safe and intuitive collaboration between humans and robots. The aim of the project is to reduce the number of unergonomic, repetitive, or physically demanding ac- tivities. 14 News 1_2020 I © DFKI The image shows robotic support for the assembly of aircraft wings, a concept being developed within the Hi-Digit Pro 4.0 project.

To guarantee the safety of the worker on such teams, project KAMeri analyzes the brain waves to deter- mine the cognitive state of the worker and conse- quently adapt the human-machine interactions. Using an easy-to-use EEG headset, brain waves are recorded and analyzed for safety-relevant states. If, for example, fatigue is detected, the robot automat- ically adjusts its activity. When the attention span drops, the robotic colleague automatically slows down its speed to prevent workplace accidents. Many complex work steps still have to be performed by human hands. This is the case, in particular, in the area of logistics such as at ports, where the variety of goods and processes makes it difficult to employ robots. The Robotics Innovation Center takes up this challenge in the KALI project, which is developing mobile, power-support systems that optimize the conditions encountered in manual tasks within the logistics processes, by reducing the physical strain and making the process more efficient. The focus is on solutions that are not only feasible, mobile, and modular, but also safe and able to gain acceptance, which involves consideration of the ethical, social, and legal issues. Intelligent platforms for intralogistics and consumer areas as well as other autonomous systems are among the research topics in Bremen because they can satisfy these requirements and support the em- ployee. A collaborative project, Knowledge4Retail, is working on a solution to automate the stocking of shelves in retail branches, which not only takes an ergonomic load off the employee, but also gives them more time for customer service. A cross-team goal of the Robotics Innovation Center is to develop an autonomous robot that can be used in areas that a high risk to humans, for example, dis- aster zones or areas requiring maintenance on ships or in tunnels. The researchers are working on sys- tems that function and perform work both in water and on land. In ﬁeld testing off the island of Hel- goland, the Bremen research team, together with the Fraunhofer Institute for Manufacturing Technology and Applied Materials (IFAM), recently tested how au- tonomous underwater vehicles can be used, for ex- ample, for the inspection and maintenance of wind turbines – work that can be life-threatening for peo- ple because of the harsh conditions that exists on the high seas. Although the importance of industrial robots is in- creasing, research does not seek to reproduce human abilities. Instead, it is focused on guarantee- ing people’s safety, protecting them from unneces- sary risk, and utilizing the workforce when and where artiﬁcial intelligence meets its limits. The Donbot robot system reaches into a shelf in a replicated retail store at the University of Bremen's Institute for Artificial Intelligence (IAI). Photo: IAI/Universität Bremen MORE INFORMATION Ħ www.dfki.de/ric CONTACT Ü Research Department Robotics Innovation Center Corporate Communications, DFKI Bremen Ö uk-hb@dfki.de 4 +49 421 17845 4180 In the KALI project, mobile, power-support systems are designed to reduce the physical strain when manually handling different items and make logistics processes more efficient. News 1_2020 I © DFKI 15

Key Technologies in the Digital Transformation – Minister President Malu Dreyer Discusses With 5G and AI Experts N “We are shaping the digital transformation. The state of Rhineland-Palatinate is making a decisive contribution to the development of the new high-performance mo- bile Internet with Kaiserslautern designated as a 5G model region,” explained Minister President Malu Dreyer, in comments made at the German language Conference “Nachhaltig Digital Unterwegs,” part of a new series of events sponsored by the state government. Around 150 interested visitors, multipliers, and experts attended the event at the German Research Center for Artiﬁcial Intelligence (DFKI) in Kaiserslautern to learn about digitalization and discuss opportunities and risks. Speakers included Prof. Antonio Krüger, DFKI CEO, DFKI “5G expert” Prof. Dr. Hans Schotten, and Prof. Dr. Katharina Zweig, of TU Kaiserslautern. “The state government is committed to a sustainable digital transition with generational fairness shaping our eorts. I am convinced that digitalization will help Rhineland-Palatinate to remain as liveable for future generations.” Potentials, visions and challenges in the context of AI & 5G – Minister President Malu Dreyer put questions to the experts Prof. Katharina Zweig (TUK) and Prof. Hans Dieter Schotten Photo: State Chancellery of Rhineland-Palatinate/ Alexander Sell The Living Labs at DFKI provide true-to-life in- formation about possible future AI applications. Projects such as the ﬁre brigade drones, 5G masts made of wood, the inter-municipal Dig- ital City network, AI-based monitoring of heart failure patients, and geodata management allow visitors to experience the opportunities in many areas of life offered by these new key technologies. The event is the second in a series of three events focusing on digitalization in all areas of work and life. It was scheduled between the Digital Forum 2019 in Mainz and the Digital Forum on 4 September 2020 in Trier. CONTACT Ü Christian Heyer Head of Corporate Communications DFKI Kaiserslautern Ö Christian.Heyer@dfki.de 4 +49 631 20575 1710 Malu Dreyer, Minister President Rhineland-Palatinate “5G networks are becoming a reality in Germany, with the ﬁrst steps here in Kaiser- slautern. Everyday life in the future will see the widespread use of smart and AI-based applications. AI and 5G will make many things in everyday life faster and easier, for ex- ample, telemedicine improves health care in rural areas,” said MP Dreyer. “We want technical innovation to lead to social progress.” Rhineland-Palatinate has very good locational advantages as a research center for developers and users. In fact, the Min- ister President added: “We are providing around EUR 18 million to strengthen AI re- search in Rhineland-Palatinate”. “5G and AI are key technologies in the digital transformation. Their integration promises completely new application possibilities for industry and society. We want to further research and sustainably develop this synergy with the Technical University and DFKI in the Kaiserslautern model region. Particularly, in new ﬁelds of technology such as 5G, it is important to identify beneﬁcial applications and to create trust through reliable research and social dialog.” Prof. Dr. Hans Dieter Schotten, Scientiﬁc Director of the Research Department Intelligent Networks at DFKI and holder of the Chair for Radio Communication and Navigation at the Technical University of Kaiserslautern (TUK) MP Malu Dreyer and Prof. Antonio Krüger in the driving simulator, where technologies for the detection of vehicle passengers are tested. 16 News 1_2020 I © DFKI

AI Research for Practical Use – DFKI Establishes Branch Office at the University of Trier N The German Research Center for Artiﬁcial Intelligence (DFKI) and the University of Trier are expanding their cooperation in ap- plication-oriented AI research. On March 16, 2020, DFKI CEO Prof. Dr. Antonio Krüger and University President Prof. Dr. Michael Jäckel signed an agreement establishing the DFKI branch office in Trier. Future joint research will focus on the intersection of artiﬁcial in- telligence and business informatics in the areas of Experience- Based Learning Systems and Cognitive Social Simulation. The new research will be performed in the university-based re- search groups of Prof. Dr. Ralph Bergmann and Prof. Dr. Ingo Timm, in cooperation with the DFKI research department Smart Data & Knowledge Services in Kaiserslautern. Strengthening the AI network and transfer In the words of Minister President Malu Dreyer: “Rhineland-Palati- nate is an innovation state. DFKI is a leader in artiﬁcial intelligence research. The branch office in Trier is a milestone for Rhineland- Palatinate and the region as we expand AI research and create a powerful network. At the same time, the topic of artiﬁcial intelli- gence has the potential to noticeably change our lives, our society, our work, and our economy. Rhineland-Palatinate sets a priority on artiﬁcial intelligence research and applications for the well-being of people, society, and the environment. Artiﬁcial intelligence can be of great service to us when it comes to mastering the urgent social challenges of our time.” “The cooperation between the University of Trier and DFKI marks a groundbreaking decision. Teaming with one of the most renowned non-university research institutes in the ﬁeld of Artiﬁcial Intelli- gence promises great advantages for all parties. Together, the two partners can proﬁtably expand their current competencies and improve the competitive position of both. Rhineland-Palatinate’s medium-sized companies also stand to beneﬁt from the transfer of knowledge, for example, from research in the ﬁelds of ‘Work 4.0’ and ‘AI in Care’.” Prof. Dr. Konrad Wolf, Minister for Science, Further Education, and Culture of the State of Rhineland-Palatinate The new DFKI branch office will be established on Campus II of the University of Trier (left). In addition to Kaiserslautern (right), DFKI maintains four other locations. Photo: Trier University/DFKI DFKI is expanding the cooperation with the Trier business informatics working groups of Prof. Ralph Bergmann (left) and Prof. Ingo Timm (right). Photo: Trier University Prof. Dr. Michael Jäckel, President of the University of Trier, is looking forward to the cooperation with DFKI: “For the computer sciences in Trier, this is another major cooperation component with non-uni- versity research institutions. In this way, we are strengthening the transfer of research results into practical, beneﬁcial artiﬁcial intelli- gence applications for important future topics.” “With the establishment of the Trier DFKI branch office, we are intensifying our long-standing contacts with the University of Trier and the working groups of our internationally renowned colleagues Bergmann and Timm. It represents an expansion of our research spectrum by two relevant AI topics and a promising potential for practical application.” Prof. Dr. Andreas Dengel, Site director DFKI Kaiserslautern and head of Smart Data & Knowledge Services research department of DFKI in Kaiserslautern Intelligent methods for the management of the future Based on combined expertise, research on how AI can be applied to learning systems aims to overcome today’s problems. “The com- bination of Case-based Reasoning approaches with Machine Learn- ing and Deep Learning methods holds enormous potential for intel- ligent process management,” said Prof. Ralph Bergmann. Application examples are ﬂexible work and production processes in trade and industry, but also decision support in terms of data protection law. Cognitive social systems research focuses on the processes for care, production, and logistics, as well as for public safety. These should be designed in such a way that the customer or service provider is actively involved in the planning and later in the execution. Prof. Timm also emphasizes the new perspectives for young scientists: “This cooperation gives our students the opportunity to work on practical AI research projects, whether as their ﬁnal thesis topic or to gain valuable experience for their own professional careers.” CONTACT Ü Prof. Dr. Andreas Dengel Site Director DFKI Kaiserslautern Head of Research Department Smart Data & Knowledge Services Ö Andreas.Dengel@dfki.de 4 +49 631 20575 1000 News 1_2020 I © DFKI 17

The speakers of the CoCoLAd-Workshop (Human-Machine Co-Creation, Co-Learning and Co-Adaptation), including Prof. Didier Stricker (2nd from left) and chair Prof. Andreas Dengel (8th from left). Photo: Inria Global AI Forum for Humanity in Paris – Responsible Development of AI N “Human-centric AI” was the theme of the Global AI Forum for Humanity (GFAIH), which took place in Paris last October and or- ganized as part of the French G7 presidency. Under the sponsorship of President Emmanuel Macron, researchers and innovation drivers from the computer, engineering, humanities, and social sciences, as well as representatives from business and politics met for three days to discuss central questions of ethics, security, and the legal status of artiﬁcial intelligence. One intention was to develop a common understanding of how the different perspectives and challenges of AI can be meaningfully understood and mastered. On this basis, a global partnership, the “Global Partnership on AI – GPAI,” is to be established and its work published in speciﬁc position papers and agendas. Prof. Dr. Andreas Dengel, one of the co-organizers of the confer- ence, said: “We know that modern AI technologies have the po- tential to assist people in many areas of life and to offer a higher degree of automation, but also that such systems if used incor- rectly can a danger. In Paris, an important step in the global net- working of experts was taken. On the basis of jointly developed principles, an AI roadmap to manage many of these issues was agreed. It became clear once again that, in addition to technology, many facets and details must be taken into account in the devel- opment, introduction, and use of AI systems. Transparency, open- ness, and trust are desirable goals in the context of AI. Keeping people in a controlling position at all times is also a priority. These goals require diverse and inter- disciplinary cooperation, and I must add, a discussion and deﬁnition of social and politi- cal goals and strategies on the basis of the current develop- ments and the state of re- search. This is the only way to achieve ‘Human-centric AI.’” The organizers met at the Élysée Palace, where Em- manuel Macron took up the pulse of the experts as he reaffirmed the importance of “Human-centric AI” meets “AI for Humanity” – Prof. Andreas Dengel and President Emmanuel Macron. 18 News 1_2020 I © DFKI President Macron reiterated the importance of developing human- centric AI systems. Photo: Inria developing human-centric AI systems and the need for interna- tional cooperation in global competition. The Global Partnership on AI was established at the last G7 Summit and led by France and Canada. This year marked the official launch and, at the end of the panel discussions, key aspects were set down in writing in a ﬁrst position paper titled “For a Global Partnership on Responsible Artiﬁcial Intelligence.” The paper highlights the pro- motion and protection of human-centric and ethical approaches to AI based on human rights; the promotion of growth, innovation, and well-being through AI; and the promotion and protection of democratic values, processes, and institutions. MORE INFORMATION Ħ https://gfaih.org CONTACT Ü Prof. Dr. Andreas Dengel Site Director DFKI Kaiserslautern Head of Research Department Smart Data & Knowledge Services Ö Andreas.Dengel@dfki.de 4 +49 631 20575 1000

Accelerating AI Together – DFKI Welcomes NVIDIA as a Shareholder NVIDIA, a world market leader in computer graphics and AI compung, has joined the DFKI shareholder group. The commitment is a further step in the long-standing successful coope- raon between NVIDIA and DFKI in the ﬁeld of machine learning and deep learning. N “NVIDIA and DFKI are working together to provide powerful tools so that scientists can solve the most complex AI challenges,” said Ian Buck, vice president and general manager of Accelerated Computing at NVIDIA. “As a DFKI shareholder, we’ll continue to jointly ex- plore how accelerated computing can beneﬁt new areas of research.” Professor Dr. Andreas Dengel, site director at DFKI Kaiserslautern and head of the Smart Data & Knowledge Services research department, stated: “The partnership between DFKI and NVIDIA combines sophisticated models and procedures with high-performance AI platforms. Having NVIDIA join as a shareholder is another building block in our long and successful cooperation, and it opens up new perspectives in application-oriented neural network research.” By combining DFKI technologies with NVIDIA’s leading AI platforms, DFKI researchers and scientists have achieved remarkable results in the application, explainability, and optimi- zation of neural networks and deep learning methods in recent years. In 2018, a team from DFKI’s Deep Learning Competence Center received the NVIDIA Pioneer Award for its work on neural network processing. For the ﬁrst time, their methods helped to shed light on the processes of the “black box” of deep learning and to render decision-making pro- cesses more understandable. High-performance infrastructure for efficient processing Several NVIDIA DGX systems in an InﬁniBand network, including the ﬁrst DGX-2 super- computer in Europe, are part of a highly efficient deep learning data center located at DFKI-Kaiserslautern. The powerful infrastructure allows for a large number of evaluations in a short time. As a result, large data volumes, such as high-resolution satellite images, can be analyzed quickly. In the same way, bottlenecks in architectures can be identiﬁed and models can be adapted accordingly – a fundamental part of engineering neural net- works. The computing power in the cluster can be optimized, idle times minimized, and deep learning methods used in a more energy-efficient manner. Photo: NVIDIA The combination of sophisticated machine learning methods and high-performance hard- ware accelerates AI applications. Making innovative technologies usable for industry DFKI researchers are also planning to make these breakthroughs usable for the industry. “We are working on complex AI algorithms to be made available for practical use in in- dustry,” said Andreas Dengel. The goal is an ‘AI Model Store’ where various machine learning methods are offered. According to plan, “A fully automated search of the archi- tectures will ﬁnd the appropriate models. Companies can select the one that best ﬁts their requirements, apply their data to the pre-trained methods, and receive feedback in real-time.” MORE INFORMATION Ħ www.nvidia.com/de-de/deep-learning-ai LINK TO “DFKI DEEP LEARNING SANDBOX” VIDEO Ħ https://youtu.be/4OeXd-XvVbo CONTACT Ü Prof. Dr. Andreas Dengel Site Director DFKI Kaiserslautern Head of Research Department Smart Data & Knowledge Services Ö Andreas.Dengel@dfki.de 4 +49 631 20575 1000 NVIDIA headquarters in Santa Clara, California. NVIDIA is the world leader in computer graphics and AI computing. Photo: NVIDIA News 1_2020 I © DFKI 19

l.-r.: Dr. Pascal Revel, Science and Technology Advisor, French Embassy; Dr. Bruno Sportisse, CEO Inria; Prof. Antonio Krüger, CEO DFKI; State Secretary Heike Raab, Representative of the Land of Rhineland-Palatinate to the Federal Government, with responsibility for Europe, the Media, and Digital Affairs; Tobias Hans, Minister President of the Saarland. Inria and DFKI Sign Memorandum of Understanding for Cooperation in AI On the ﬁrst anniversary of the Aachen Treaty, the French Naonal Instute for Informaon Technology (Inria), and the German Research Center for Arﬁcial Intelligence (DFKI), signed a Memorandum of Understanding. The ceremony took place in Saarbrücken at the Power4Pro- ducon innovaon laboratory in the presence of Tobias Hans, Minister President of Saarland and State Secretary Heike Raab, Representave of the Land of Rhineland Palanate to the Federal Government, with responsibility for Europe, the Media, and Digital Aﬀairs. N The Aachen Treaty of January 22, 2019, between France and Ger- many, has given new impetus to Franco-German cooperation and also highlighted the partnership in the ﬁeld of artiﬁcial intelligence (AI). Under this framework, Inria and DFKI have agreed to signiﬁ- cantly strengthen their collaboration in the ﬁeld of AI by structuring and formalizing their cooperation. This agreement is documented in the Memorandum of Understanding between Inria and DFKI sig- ned in January 2020. Minister President Tobias Hans stated: “DFKI and Inria have develo- ped a common vision for the opportunities and challenges posed by artiﬁcial intelligence and digital sovereignty in Europe. The close cooperation of both internationally renowned research institutes is a groundbreaking step towards making Artiﬁcial Intelligence in the Saarland and France a driving force for future technologies in Europe. This agreement also shows that Saarland is uniquely at- tractive and enjoys a signiﬁcant standing as a center for IT research. The aim of our strategy to increase cooperation with our French partners is bearing fruit.” State Secretary Heike Raab added: “Germany and France are focu- sing on people in artiﬁcial intelligence R&D. Human-centric AI is a central aspect of this common vision of values. I am very pleased that cooperation between DFKI and Inria, as two internationally sig- niﬁcant research institutions, will further intensify with the signing of this MoU. For the Kaiserslautern site, this speciﬁcally means that cooperation in the areas of machine learning and deep learning will be expanded. DFKI is already recognized internationally in these areas of expertise and this cooperation will only serve to enhance that.” 20 News 1_2020 I © DFKI

• Applying AI to Sustainability (e.g., Climate) • AI to improve the competitiveness of European industry • AI for education and training The partners are conﬁdent that cooperation on these issues will help build a society that unites the people of Europe, that includes everyone, and uses technology to serve the needs of the people. In the future, Inria and DFKI plan to work together to advance re- search programs and activities on a European level in a coordinated manner. Joint proposals have already been submitted under the ICT 48 call for the creation of AI Centers of Excellence. A joint stee- ring committee consisting of representatives of the management of DFKI and Inria will deﬁne concrete projects for cooperation and support their implementation. MORE INFORMATION Ħ www.inria.fr Ħ https://claire-ai.org CONTACT DFKI Ü Dr. Anselm Blocher Representative for German-French activities at DFKI Ö Anselm.Blocher@dfki.de 4 +49 681 85775 5262 CONTACT INRIA Ü Laurence Goussu Inria – Institut National de Recherche en Sciences et Technologies du numérique Ö Laurence.Goussu@inria.fr 4 +33 6 81 44 17 33 “This MoU between DFKI and Inria illustrates the momentum of the French-German coopera- tion on AI within the European dynamic following the Aachen Treaty and the recent meeting in Toulouse. Our national plans regarding AI promote bilateral partnerships between research organizations from both countries. DFKI and Inria share a similar and open vision for AI and this paves the way to strengthen our partnership and create new project opportunities, both at bilateral and European levels.” Bruno Sportisse, Inria CEO DFKI and Inria have a long history of scientiﬁc cooperation and col- laboration. This exchange was strengthened in 2019 through bila- teral visits and regular high-level meetings that have led to a shared vision for the challenges of AI and digital sovereignty in Europe. “This agreement documents our eorts not only in the area of French-German AI cooperation, but also in actively coordinating our research and deﬁning joint project goals. We look forward to this collaboration as we bring together the French and German visions of AI for the beneﬁt of people and society. Both countries have initiated pro- grams that ﬁll the term human-centric AI with content and create value for France and Germany and European society as a whole.” Antonio Krüger, DFKI CEO In the future, the two research institutes will cooperate within the framework of a joint strategic research and innovation agenda in the areas of AI for INDUSTRIE 4.0, AI infrastructures, AI and cyberse- curity, human-robot cooperation and wearables, and additional to- pics. At the heart of the agreement is a strong joint commitment to the European AI initiative CLAIRE (Confederation of Labs for AI Research in Europe), which aims to bring European AI researchers closer together and to jointly advance European AI research that beneﬁts people while respecting fundamental European values. Speciﬁc measures include the implementation of joint research and innovation projects, which partially build on existing projects at both organizations, as well as some new topics, which are deﬁned and worked out in joint workshops. The signing of the MoU was pre- ceded by a workshop with more than 80 participants in Nancy, France on January 20-21. Both institutes explored common poten- tials as the prelude to regular working group meetings and ex- changes. AI applications on which these projects will focus are: • Trust through Explainability and Certiﬁcation of AI systems • Overcoming the language barriers in Europe and the world • Breakthroughs in AI for Health and Personalized Medicine • Better HW/SW infrastructure for AI (e.g., TPUs, FPGA, • Quantum, etc.) News 1_2020 I © DFKI 21

CLAIRE Receives Broad Mandate and Funding to Shape “AI Made in Europe” The European Commission has taken a signiﬁcant step towards strengthening AI research in Europe by allocang EUR 50 million of seed funding to prepare the ground for much more substanal investments in the near future. This is crically necessary to keep Europe compeve with countries such as the USA, China, and Canada, which are invesng substanally higher amounts into AI research and innovaon. N Four proposals coordinated by members of CLAIRE A total of ﬁve research proposals will receive funding under the ICT- 48-2020 call for proposals, four of which will be led by members of CLAIRE (Confederation of Laboratories for Artiﬁcial Intelligence Re- search in Europe), the world's largest network of AI research groups and institutions. Three of these projects are TAILOR, Humane-AI-Net, and AI4Media. The VISION network will coordinate between these and another network selected for funding to position Europe as a leader in a human-centric, trustworthy AI. In this highly competitive process, CLAIRE once again consolidates its position in shaping the Euro- pean AI ecosystem. TAILOR – Network of research centers for foundations of trustworthy AI The Europe-wide network TAILOR under the direction of Prof. Fredrik Heintz from the University of Linköping, Sweden, is devel- oping a research and innovation roadmap for trustworthy AI with special emphasis on the integration of machine learning, case- based reasoning, and combinatorial optimization. TAILOR will prepare the ground for AI research that addresses the major chal- lenges in the ﬁelds of health, mobility, and resource management. HumanE-AI-Net – Research on AI systems that enhance human intelligence HumanE-AI-Net also includes members of the ELLIS (European Lab- oratory for Learning and Intelligent Systems) network. Prof. Paul Lukowicz, head of the DFKI Embedded Intelligence research de- partment, is the lead coordinator. HumanE-AI-Net will develop the scientiﬁc and technological foundations necessary to build AI sys- tems that enhance rather than replace human intelligence. (www.humane-ai.eu) AI4Media – AI for media and society AI4Media, under the leadership of Dr. Yiannis Kompatsiaris, Senior Researcher at the Center for Research and Technology Hellas (Greece) and supporter of CLAIRE, will establish a network of cen- ters of research excellence with a focus on the media sector. VISION – Coordination across Europe VISION coordinates among the new network of AI centers of excel- lence and with the European Commission. It is headed by Prof. Holger Hoos from the University of Leiden in the Netherlands, who also chairs the Executive Board of CLAIRE. Marc Schoenauer, Deputy Research Director at the French research institute Inria, is a key contributor to the French AI strategy and is also involved in VISION. At the beginning of the year, Inria and DFKI signed a Memorandum of Understanding to signiﬁcantly enhance their long-standing cooperation. EU White Paper supports CLAIRE vision The European Commission outlined a plan on February 19, 2020, that addresses all key elements of CLAIRE's vision for “Excellence in AI.” The success of the CLAIRE network in ICT-48-2020 conﬁrmed its pioneering role in deﬁning and implementing the European vi- sion for excellence in AI research and innovation. The Commission strongly argues that “the EU must act as one.” This is directed not only to the member states but also to the Euro- pean AI research community. Rather than highlighting differences in methodology and techniques, it is essential to focus on “excel- lence across the AI community” by bringing the players together and pooling their expertise to realize the many promises of AI for Europe. This part of the CLAIRE vision is now strongly supported by a recent White Paper. Photo: Pete Lomchid, Getty Images 22 News 1_2020 I © DFKI

Ursula von der Leyen, President of the European Commission, in conversation with Hans De Canck, head of the AI Experience Center and the CLAIRE office in Brussels. Photo: CLAIRE The White Paper recognizes the need for a broad network of AI research centers and larger institutions to focus on strategic appli- cations, ranging from agriculture to energy and ﬁnancial services. More importantly, it calls for the creation of a European AI hub – “a center of AI research and innovation” to attract talent from around the world. “The concept of a hub, inspired by the widely recognized success of CERN, has been one of the key elements in the CLAIRE vision since we ﬁrst presented it in June 2018. This lighthouse project will bring together top researchers from all over Europe and the world. It will become a focal point for large-scale research initiatives (e.g., in robotics or autonomous transport) and a widely seen as a visible symbol of European excellence and ambition in the ﬁeld of artiﬁcial intelligence.” Prof. Philipp Slusallek, Scientiﬁc Director at DFKI and co-initiator of CLAIRE s In June 2018, hundreds of Artiﬁcial Intelligence experts from across Europe founded the CLAIRE research network to capitalize on ideas and insights on how Europe can take a leading role in Artiﬁcial Intelligence research and innovation. Since then, CLAIRE has grown to more than 350 AI research laboratories and organizations representing more than 21,000 researchers and staff from across Europe. MORE INFORMATION Ħ https://claire-ai.org CONTACT Ü Prof. Dr.-Ing. Philipp Slusallek Site Director DFKI Saarbrücken CLAIRE Co-Initiator Ö Philipp.Slusallek@dfki.de 4 +49 681 85775 5377 News 1_2020 I © DFKI 23

our Data Analytics unit. We are excited to have found DFKI, one of the most important international centers of excellence with more than 1,000 researchers for AI and Deep Learning. Our aim is to strongly expand this business activity with strong partners joining us in the de- velopment of such solutions,” said CEO Dr. Joachim Kreuzburg. Sartorius and DFKI develop and employ image and pat- tern recognition processes, machine learning, and other technologies for life science applications. For example, researchers are working on new deep learning algorithms and methods for use in image recognition of cells and or- ganoids, analysis and modeling of biological systems, and simulation and optimization of biopharmaceutical pro- duction processes. Sartorius provides extensive data for this purpose and sends specialists from its corporate re- search and product development units to SAIL. Prof. Andreas Dengel, executive management board member at DFKI and head of the Smart Data & Knowledge Services research department at Kaiserslautern said: “The DFKI transfer labs develop new service modules with arti- ﬁcial intelligence for various markets and target groups. We have been using our AI hardware equipment, which is unique in Europe, for the past two years in application-ori- ented research together with renowned companies such as Allianz, Continental, Hitachi and IAV. I am very pleased that we are entering the life science tools market as a partner with a global player like Sartorius.” MORE INFORMATION Ħ www.sartorius.com CONTACT Ü Prof. Dr. Andreas Dengel Head of Research Department Smart Data & Knowledge Services Ö Andreas.Dengel@dfki.de 4 +49 631 20575 1000 The Sartorius Campus in Göttingen. Photo: Sartorius Sartorius and DFKI Launch Joint Research Lab DFKI has partnered with Sartorius, a leader in the biopharmaceucal industry and internaonal life science research, to establish the Sartorius AI Research Lab (SAIL). The use of arﬁcial intelligence (AI) in Sartorius products and sooware pla|orms is being experimentally tested and further developed at SAIL on the DFKI-Kaiserslautern campus. N SAIL is assigned to the DFKI Smart Data & Knowledge Services research de- partment headed by Professor Andreas Dengel. The department uses DFKI’s deep learning hardware and expertise, which is unique throughout Europe. As a protected data room and independent data laboratory, SAIL is open to Sar- torius partners and customers as part of cooperation agreements, and the ﬁrst projects have already been initiated. Sartorius and DFKI will use SAIL for training purposes and DFKI employees will be able to further their education in life sci- ence applications at Sartorius research sites. “Advanced methods of data analysis have been used only to a very limited ex- tent in the biopharmaceutical industry, for drug research and production. This is one of the reasons that development timelines and costs for medical drugs are forever increasing. Better use of data through AI is a particularly promising approach for making signiﬁcant progress in this area. Many customers are al- ready using our powerful software for speciﬁc applications in biopharmaceu- tical production. We began offering the product two and a half years through Prof. Oscar-Werner Reif (2nd from left), Chief Technology Officer at Sartorius, visiting the laboratory at DFKI-Kaiserslautern. 24 News 1_2020 I © DFKI

Walkers support independent mobility. Photo: TOPRO GmbH Safer Walker Prevents Falls Thanks to Intelligent Posture Recognition Many people ﬁnd the use of a walker to be a prerequisite for an acve and independent life – but, using one may also be dangerous. The DFKI research department Cyber- Physical Systems is part of the ModESt project, a consorum that is developing an in- telligent support system that can prevent falls involving the use of a walker. N The aim is to effectively prevent latent postural mistakes and the related risk of falling down. In the project, a new type of electronic box is integrated into the wal- ker. The box links distance sensors for posture recognition with software-based al- gorithms to identify incorrect postures and provide feedback modules that display instructions for the corresponding corrective actions. The discreet feedback of pro- blem postures is important, in particular, for people who are unable to internalize correct posture over the long term due to cognitive limitations. DFKI's Cyber-Physical Systems department, headed by Prof. Dr. Rolf Drechsler, ﬁrst created the technical prerequisites for recording posture patterns. The researchers determined the optimal positions for the distance sensors on the frame of the walker and developed software for evaluating posture patterns. Additio- nally, they developed a concept to adjust the walker to the individual in the best possible manner. Overall, the treating physicians can analyze the progress of therapies for regaining mobility in a sound manner under strict data protection criteria. The electronic box forms a reliable, cost-effective, and innovative concept for correcting physical mal- positioning when using the walker and contributes to creating real added value in the everyday mobility of many people. The collaborative project is sponsored by the Federal Ministry of Education and Research (BMBF) as part of the program “Human-Machine-Interaction in the shifting demographic.” The consortium includes four The electronics box con- nects data from distance sensors on the walker with software algorithms to iden- tify faulty postures. partners from science, health, and industry. In ad- dition to project coordinator Budelmann Elektronik in Münster and DFKI Bremen's Cyber-Physical Sys- tems research department, the remaining partners are GesundheitNord Klinikverbund in Bremen, with its specialist department for geriatrics, and TOPRO company located near Munich. MORE INFORMATION Ħ www.dfki.de/cps CONTACT Ü Dr. Serge Autexier Research Department Cyber-Physical Systems Ö Serge.Autexier@dfki.de 4 +49 421 218 59834 News 1_2020 I © DFKI 25

Intelligent Environmental Protection and Smart Services from DFKI Laboratory Niedersachsen Using drones to locate plasc refuse, digitalizing the agricultural and hou- sing sectors, proacvely combang wear and tear in manufacturing ma- chinery – at ﬁrst glance, these projects at DFKI Laboratory Nieder- sachsen seem to have lile in common. However, they all pursue a com- mon goal: to place a human-centric focus on research in applicaon-ori- ented development and sustainable sooware. This young lab, roughly one year aoer its establishment, is showing what digital technologies can con- tribute to comprehensive environmental protecon and how important this development is for Lower Saxony and beyond. N In April 2019, DFKI Laboratory Niedersachsen was founded in Osnabrück and Oldenburg for the purpose of promoting the use of AI in agriculture and marine management as well as in advanced production systems, all the while building on regional structures. Thanks to strong local partners, close cooperation with the state government, and pioneering, application-oriented projects, the value the laboratory produced in the ﬁrst year exceeds the ex- pected budget volume for the third year mark. The three research areas of Plan-Based Robot Controls (PBR), Ma- rine Perception (MAP), and Smart Enterprise Engineering (SEE) cover a wide range of AI application areas – from autonomous robot systems in agriculture and transportation to intelligent sensors for aquatic environments to smart products and new business models. The overarching theme is the same: acquisition, processing, and value-adding use of sensor data. It involves geo-data from satellites as well as process data from complex machines, and contextual knowledge about physical and virtual environments of agents and robots. An essential building block for robust and justiﬁable AI is the ability to perceive the sur- roundings, and this poses numerous issues that are linked to current challenges, for example, the development of central cloud solutions that provide the data collected. A ﬁrst joint project at DFKI Labora- tory Niedersachsen is currently researching the topic of dynamic symbol anchoring, which should lead to a fundamental improve- ment in human-machine interaction. A researcher from DFKI Laboratory Niedersachsen tracks the route of the Arox robot system on a field from a tablet. The processing of sensor data also plays a major role in environ- mental protection, for example, as it relates to the question of how digitalization can promote more sustainable agriculture. In a global context, intelligent sensor technology can be used to combat plas- tic waste: A project in the MAP research area is equipping drones with technologies that can detect and track waste accumulations in rivers and oceans. MORE INFORMATION Ħ www.dfki.de CONTACT Ü Prof. Dr. Joachim Hertzberg Spokesperson, DFKI Laboratory Niedersachsen Ö Joachim.Hertzberg@dfki.de 4 +49 541 386050 2251 (l.-r.) Patrick Rosen (ROSEN Group), Prof. Oliver Zielinski (MAP), Prof. Susanne Menzel-Riedl (President of Osnabrück University), Stephan Weil (Minister President of Lower Saxony), Prof. Oliver Thomas (SEE) and Prof. Joachim Hertzberg (PBR) at the reception at DFKI Laboratory Niedersachsen in the city of Osnabrück. k c o t S e b o d A : o t o h P 26 News 1_2020 I © DFKI

An AI algorithm enables the predictive evaluation of traffic situations. LiDARshared – From the Networked Bus to Safe, Autonomous Use in Shared Space Photo: Blickfeld GmbH N “Shared Space“ is a traffic control concept in which all road users, whether cars, rental bikes, e-scooters, or pedestrians, have equal rights. This concept essentially has no traffic signs, signal sys- tems, or road markings. The Shared Space concept offers an im- proved quality of mobility for the public, but autonomous driving has its challenges: There are many more paths that a road user may take in the next minute, which demands a huge increase in the number of planning alternatives. The most important technology for accident prevention is LiDAR (Light Detection and Ranging). The radar-related method for dis- tance and speed measurement uses laser beams instead of radio waves to visually detect obstacles in driver assistance systems and for autonomous driving. The recently launched LiDARshared project is developing AI-based systems for safety and accident prevention involving autonomous vehicles using LiDAR sensors in the “shared space.” The project is trying to network environmental data in real-time from multiple involved vehicles. The major elements of the project are the pro- vision of open data via the central access point mCloud of the Fe- deral Ministry of Transport and Digital Infrastructure (BMVI) and the subsequent evaluation using AI methods. The Transportation Management Company of Kiel, Germany is in- tegrating the system into its buses and will be demonstrating it at major events such as the international sailing event “Kieler Woche”. To have the software installed in the buses and to use networked LiDAR data in real-time during such an event makes adaptation to other application scenarios much easier while also enabling its transfer to individual transport systems. The technology developed in LiDARshared improves the safety of autonomous vehicles and helps to intelligently control traffic ﬂow in the future. A broad range of innovative applications is made possible for cities, municipalities, and the environment through the use of networking and AI-enhan- ced data. The project leader is Blickfeld, a supplier of 3D LiDAR pro- ducts. Other project partners are ADDIX Internet Services, DFKI, and KVG (the public transport company of Kiel, Germany). LiDARshared is funded by BMVI with almost EUR 1.6 million for a period of two years. The Agents and Simulated Reality research department at DFKI is re- sponsible for the development and training of an AI algorithm for the predictive evaluation of the traffic situation. The movement trajecto- ries of all road users in “shared space” are predicted on the basis of data from the simulation environment. Using the sensor data from several LiDAR, cameras, and GPS, the algorithm enables a rolling, ﬁve-second forecast of road user movement patterns. MORE INFORMATION Ħ www.lidarshared.de Ħ https://mcloud.de CONTACT Ü Dr. Christian Müller Head of DFKI Competence Center for Autonomous Driving Ö Christian.Mueller@dfki.de 4 +49 681 85775 4823 on the basis of a decision by the German Bundestag Supported by: News 1_2020 I © DFKI 27

Sub-symbolic AI methods – neural networks, deep learning, machine learning, etc. are achieving breakthroughs that were unimaginable just a few years ago. We are now able to analyze vast volumes of data on which to base conclusive statements: The areas of applicaon include obstacle detecon in autonomous driving, forest damage assessment, and tumor detecon. AI tools are already used to ﬁnd suitable partners on online dang pla|orms and partner search engines, and company personnel managers use them to select the best applicants. Even though Arﬁcial Intelligence is not making the decision or acng on its own, we sll have to understand why the vehicle's onboard system recognizes an object as an obstacle, or why an algorithm recommends one applicant and rejects another. Otherwise, deﬁciencies could persist, and prejudices become evaluaon criteria. An interdepartmental eﬀort is underway at DFKI to make deep learning methods explainable by trying alter- nave approaches. Two of these are based in the Smart Data & Knowledge Services research department and are presented below – Project ExAID and a new visualizaon framework for me series analyses: N Explainable AI Diagnostics for early identiﬁcation of complex illnesses – ExAID An explainable diagnosis assistance system for physicians The diagnostic mode offers practicing physici- ans an intelligent assistance system that provi- des diagnostic recommendations and substan- tiates them with conceptual, visual, and textual justiﬁcations. Training tool for physicians & analysis tool for AI experts The system also offers a training and analysis mode that provides intuitive insights into the distribution of diagnostic criteria across entire data sets for aspiring as well as experienced dermatologists. This mode also enables AI de- velopers to carry out an in-depth analysis of the decision-making processes in order to validate system functions, identify undesirable inﬂuen- ces, and, if necessary, initiate corrective mea- sures. The introduction of such diagnostic assistance systems in everyday clinical practice signiﬁ- cantly reduces the workload and frees medical professionals of routine tasks, which means that there is more time for research and compre- hensive patient care. In the future, besides ad- ditional use case scenarios in dermatology, such systems will expand into other ﬁelds. In light of the aging global population and a continuing shortage of skilled workers worldwide in the healthcare sector, the early detection of disease cannot happen soon enough, for example, in diagnosing skin cancer. Recent developments and successes in the area of deep learning have produced a large number of computer-aided diag- nostic systems that can match the performance of doctors and, in some cases, even exceed them. However, the practical use of such “black box” systems is being held back by a lack of acceptance and ethical and legal concerns. That highlights the urgent need to de-mystify and explain such systems. Concept distribution Developers Diagnosis is s y l a n A Concept inﬂuence Concept presence D i a g n o sis AI Diagnosis Concept region Physicians Researchers ExAID explains a diagnosis assistance system for medical professionals. DFKI, in cooperation with dermatologists of the University Hospital Münster, is presenting a diagnostic assistance system that serves as an understandable interface between physicians and AI algorithms. The project examines existing AI-based diagnostic sys- tems in terms of the diagnostic criteria used in decision support systems. The diagnosis criteria are abstract concepts deﬁned by physicians, which help them in making a di- agnosis. The system offers two different functional modes: 28 News 1_2020 I © DFKI

N Visualization Framework for Time Series Analysis – TSViZ Deep neural networks (Deep Neural Networks, DNNs) are used in a number of different applicati- ons, such as for searching or for recommendations and automated analysis. The issues of reliability and robustness are important. For critical safety areas, not only these aspects but also accuracy and trustworthiness are extremely important. DFKI is developing the Time Series Visualization Framework (TSViZ) to interpret DNNs and provide appropriate information about the decisions of neural networks for time series analysis. The infor- mation is useful, for example, in forecasting wear and tear and maintenance services. The system acts as an interface between researchers, devel- opers, and users to explain the how and why of a particular decision made by the neural network. The framework enables different user groups to un- derstand the network predictions as well as the speciﬁc parts of the models. The insights being gained then allow for the further optimization of the framework. The users are only shown informa- tion that is tailored to their particular situation. Explaining a particular network decision is a chal- lenging task that involves several factors, such as the domain, the target group, and the data format. When it comes to the domain, most of the existing explanations are designed for pictures or speech, because people can immediately understand these modalities. This is not the case with time se- ries, which a signiﬁcant challenge because the data for the time series modality is not understan- dable through images and text. The visualization framework is trying to solve this challenge. Coverage of the time series domain The framework acts as an instance-based post- hoc procedure and can manage any CNN archi- tecture of a neural network. Furthermore, it is not limited to a speciﬁc task and can be used for clas- siﬁcation and forecasts as well as for regressions. User-speciﬁc explanations enable broad application Data that is not relevant to the background of the target group can be hidden using the interactive interface. The system is designed to be used by de- velopers and users without restriction to a speciﬁc group. Input Output Input relevance Filter output Network groups Detailed output Network optimization Optimized DNN A Time Series Visualization Framework – TSViZ – explains the how and why of a deep neural network decision. MORE INFORMATION Ħ www.dfki.de/sds CONTACT Ü Prof. Dr. Andreas Dengel Head of Research Department Smart Data & Knowledge Services Ö Andreas.Dengel@dfki.de 4 +49 631 20575 1000 CONTACT Ü Dr. Sheraz Ahmed Research Department Smart Data & Knowledge Services Ö Sheraz.Ahmed@dfki.de 4 +49 631 20575 4818 News 1_2020 I © DFKI 29

Software Innovation for the Digital Transformation – Project MOSIM in the ITEA3 Program The ongoing digital transformaon presents opportunies for faster implementa- on of new ideas for products and services, for a more ﬂexible adaptaon of the manufacturing processes, and for modiﬁed product por|olios or new prod- uct opons. The development of innovave sooware that supports the digital transformaon is the focus of the internaonal research, development, and innovaon program known as ITEA 3. ITEA is a cross-border, industry-sponsored program that is part of the European EUREKA Ini- ave, a worldwide network from 40 countries. N DFKI's Agents and Simulated Reality lab, headed by Prof. Dr. Philipp Slusallek, became part of the ITEA program with the launch of the MOSIM project. Project partners from Finland, Sweden, Austria, and Germany are working on simulating the ways people move for the automotive industry under the direction of the Daimler Corporation. Using AI methods, the MOSIM project develops software approaches to signiﬁcantly increase the degree of reality in simulations. The project uses the results of previous ITEA projects and develops them further in the automobile manufacturing environment with the goal of achieving signiﬁcant advances in terms of feasibility, er- gonomics, and time. The aim is to develop a concept based on the FMI standard (Functional Mock- up Interface). This concept applies the co-simulation method of using models from various simulations and the introduction of Motion Model Units (MMU) that facilitate transfer to human mo- tion. Projects are supported that develop innovative solutions that help companies and their customers to master the digital transition and its major challenges in ways that promise the advancement of our society. The main focus will be on companies and their customers. MORE INFORMATION Ħ https://itea3.org Ħ https://mosim.eu CONTACT Ü Dr. Klaus Fischer Research Department Agents and Simulated Reality Ö Klaus.Fischer@dfki.de 4 +49 681 85775 3917 AI and Data Analytics for Safer and More Reliable Mobility The purpose of the DAYSTREAM project is to analyze data in the BMVI mCloud and MDM in combinaon with other public source data from individual modes of transport as well as data from other mobility-related companies. The project aims at: • Developing new application prototypes for early detection, timely tracking, in-depth analysis and the best possible prediction of security and process-critical events, • Creating new data resources through the automatic interpretation, extension and semantic networking of existing data, and • Feeding the new data back into the mCloud in such a way as to increase the value of existing data and to form the basis for a variety of future applications. Localization of events from social media data. 30 News 1_2020 I © DFKI DFKI is the lead manager of the collaborative project, which includes partner companies DB Safety, idalab, Rhine-Main Transportation Service Association, and also the University of Kassel's Department of Transport Engi- neering and Logistics. A software demonstrator shows the results to date from the project, which has a planned close out in July 2020. Various data sources are linked and visualized for the analysis and improvement of the public transport ser- vices of the Rhine-Main Transport Association to include temporal and spatial evaluation of the accident risk from wild animals in the state of Hessia. The DAYSTREAM project is part of the “mFund” research initiative, sponsored by the Federal Ministry of Trans- portation and Digital Infrastructure (BMVI), (funding grant: 19F2031A). MORE INFORMATION Ħ www.daystream-projekt.de CONTACT Ü Holmer Hemsen, Ph.D. Research Department Intelligent Analytics for Massive Data Ö Holmer.Hemsen@dfki.de 4 +49 30 23895 1835 on the basis of a decision by the German Bundestag Supported by:

s u k o F r e f o h n u a r F : o t o h P QURATOR 2020 – From Experimental Concept to Routine Business The QURATOR Conference 2020 brought together representaves from the areas of culture, healthcare, logiscs, energy, industry, and the media. Lectures and workshops provided insights into digital informaon management on the basis of Arﬁcial Intelligence. N Many people on hearing the term Artiﬁcial Intelligence, accor- ding to Georg Rehm, imagine terrifying humanoid robots from the world of science ﬁction. However, as a researcher at DFKI Berlin working in the department of Speech and Language Technology, he notes a positive shift in this perception: “In particular, those re- presenting the industry are now much better able to assess artiﬁ- cial intelligence from a technical perspective.” QURATOR is an acronym for “Curation Technologies.” Two hundred ﬁfty participants attended the QURATOR Conference 2020 in Ja- nuary in Berlin. Dr. Georg Rehm is the principal researcher for this “growth center,” which is sponsored by the Federal Ministry of Edu- cation and Research (BMBF). The term curation no longer needs as much explanation as it did a few years ago, said spokesperson Armin Berger, owner of 3pc agency located in Berlin. The agency has worked with “digital curation technologies” (DKT) for many years. The name stems from one of the research initiatives in the BMBF's “Innovative regional growth centers / Potential GC” pro- gram. The project has resulted in new tools and curating services for combining, merging, and internationalizing knowledge content for research, enrichment, and analysis. QURATOR develops intelli- gent methods and procedures for the management and structu- ring of digital content. The smart newsroom The participants presented their application-oriented IT solutions and prototypes in lectures and workshops – for example, storytel- ling and exhibition curating, process and knowledge management, personalized medicine and, last but not least, the production of media content. The intelligent newsroom, for example, is beco- ming increasingly important in the age of the Internet and social media. Recent examples of media coverage discussed at the con- ference were the forest ﬁres in Australia and the shooting down of the Ukrainian passenger plane. These examples clearly show the dis- covery and compiling of facts and content in the shortest possible time is just as important as the rapid check of their accuracy or au- thenticity. Indeed, the media branch is placing high expectations on intelligent technologies. The participants at the 2-day conference were impressed by the wide range of AI solutions already available – and they conﬁrmed the notion that developers and users must work together. MORE INFORMATION Ħ https://qurator.ai CONTACT Ü Dr. Georg Rehm Research Department Speech and Language Technology Ö Georg.Rehm@dfki.de 4 +49 30 23895 1833 News 1_2020 I © DFKI 31

Professor Wahlster and Four DFKI Alumni Honored as Inﬂuential Thinkers in German AI Research At the beginning of the year, the German Informacs Society (GI) and the Fraunhofer IUK Technology Group recognized ten inﬂuenal ﬁgures in the German AI landscape. N Five prize winners have worked at DFKI at some point in their ca- reers: Besides Prof. Dr. Wolfgang Wahlster, Chief Executive Advisor and former Chairman of the DFKI Executive Board, the winners in- clude Prof. Dr. Elisabeth André, Prof. Dr. Susanne Biundo-Stephan, Prof. Dr. Bernhard Nebel and Prof. Dr. Jörg Siekmann. “Being honored as one of the ten most inﬂuential ﬁgures in AI re- search in Germany by our most important national professional as- sociations, gives me an obligation to continue to work intensively on behalf of German AI research,” said Prof. Wahlster. “It is impres- sive proof of DFKI's leading position that ﬁve of the ten personalities on this list are DFKI alumni: Elisabeth André and Bernhard Nebel, two of my former DFKI doctoral students, Susanne Biundo-Stephan, an outstanding member of my DFKI research department, and Jörg Siekmann, co-founder of DFKI and long-standing DFKI department head.” Wolfgang Wahlster was one of the initiators of DFKI, which was for- merly established in 1988 and is now Germany's leading, commer- cially-oriented research institute with over 1,000 employees. Elisabeth André, Susanne Biundo-Stephan, and Bernhard Nebel wor- ked at DFKI, mainly in the 1990s. As specialists in the ﬁelds of human-machine-interaction, plan-based AI, computer-assisted proof, cognitive systems, knowledge modeling, and knowledge re- presentation, they have contributed signiﬁcantly as project mana- gers and senior researchers to the shaping of DFKI and its research proﬁle. Prof. Dr. Jörg Siekmann was Germany's ﬁrst Professor for Artiﬁcial Intelligence in 1983 and one of DFKI's ﬁrst directors. His research in- terests are computer-assisted mathematical proofs, uniﬁcation theory, e-Learning for mathematics, formal methods, and multi- agent systems. The German Informatics Society created Project #KI50: “Artiﬁcial Intelligence in Germany – yesterday, today, and tomorrow.” In May 2019, the “Artiﬁcial Intelligence” department organized a jury com- posed of 18 experts in transdisciplinary AI, which selected ten inﬂu- ential thinkers appearing in Germany's AI history. The outstanding scientists were honored in a ceremony by the German Informatics Society and Fraunhofer IUK Technology Group at the New Year's re- ception. The greats of German AI research also met on stage with #KI50 newcomers and talked about relevant developments in the transdisciplinary ﬁeld of research. The project is funded by the Fe- deral Ministry of Education and Research (BMBF). CONTACT Ü Reinhard Karger Corporate Spokesperson Ö Reinhard.Karger@dfki.de 4 +49 681 85775 5253 Influential AI Thinkers (l.-r.): Bernhard Nebel, Wolfgang Wahlster, Theresa Tran (AI-Newcomer), Elisabeth André, Susanne Biundo-Stephan, Stefan Wrobel, Wolfgang Bibel, Gunay Kazimzade (AI-Newcomer), Christine Regitz (Chair of Jury), Otthein Herzog, Dieter Fellner (Jury member), Hannes Federrath (Jury member), Christoph Beierle (Jury member) - kneeling: Jörg Siekmann, Bernhard Schölkopf. Photo: GI/Fabian Sperk 32 News 1_2020 I © DFKI

Photo: Pixabay AI for Resilient Production – Project “SPAICER” Kick-o Event The Smart Servive Engineering research department, under the direcon of Prof. Dr. Wolfgang Maaß, is responsible for coordinang the project, which is funded unl March 2023 and centers on the subject of resilience management. Researchers are working with manufacturing companies to study how AI methods can be applied in the industry to be beer prepared for interrupons and have adequate responses available. N For many years, the leading business risk in our globalized and networked economy has been the interruption of production, in- cluding the interruption of supply chains. A company's ability to adapt to internal and external changes and disruptions is called the “search for resilience.” Reinforced by the signiﬁcant increase in pro- duction complexity as a result of INDUSTRIE 4.0, resilience man- agement has become an indispensable success factor for industrial manufacturing companies. The SPAICER project uses leading AI technologies and INDUSTRIE 4.0 standards to develop a data-driven ecosystem based on life-long, collaborative, and low-threshold Smart Resilience Services. The aim of the project is to anticipate disruptions (anticipation) and optimize production planning to manage the active disruptions at any time (response). A virtual kick-off event was organized on April 28, 2020 that allowed the project consortium to come together remotely for the ﬁrst time in its entirety. The tone for the project was set through keynote speeches and interactive workshop discussions of the ﬁrst steps. Representatives of the Federal Ministry for Economic Affairs and Energy (BMWi) and the project sponsor German Aerospace Center (DLR-PT) took part in the kick-off via a live video link. In addition to DFKI as the consortium leader, the project partners include: WZL (Laboratory of Machine Tools and Production Engi- neering) at RWTH Aachen, the University of Freiburg, the University of Darmstadt, the Institute for Technology and Innovation Mana- gement at RWTH Aachen, the Otto Beisheim School of Manage- ment (WHU), deZem, Feintool, SAP, SCHOTT, SCHAEFFLER, SEITEC, SENSEERING, and Waelzholz. SPAICER also enjoys a large and wide- ranging network of over 40 associated partners who provide project support to the consortium and represent an outstanding example of a complete value-adding network. SPAICER is partially funded by the BMWi. MORE INFORMATION Ħ www.spaicer.de CONTACT Ü Prof. Dr. Wolfgang Maaß Head of Research Department Smart Service Engineering Ö Wolfgang.Maass@dfki.de 4 +49 681 302 64736 R E C A P S I : o t o h P News 1_2020 I © DFKI 33

Artiﬁcial Intelligence in Networked Production – Project AIRPoRT Innovave informaon and communicaon technologies are one of the main elements driving today's development of producon environments. At the communicaon level, they enable ever faster wireless data links, for example, through the use of 5G in the control and coordinaon of machines and mobile robots. In terms of data and informaon processing, AI and machine learning processes support the eﬃcient and eﬀecve use of very large amounts of data. The opmizaon of exisng individual producon processes is not all: technical and organizaonal innovaons all along the enre process chain are also possible. N Project AIRPoRT (Artiﬁcial Intelligence for Robotics and Networked Production) de- velops, optimizes, and tests these technologies. The cooperation partners contribute their speciﬁc focus and experience: The DFKI research department of Smart Data & Knowledge Services develops innovative AI-based methods for analyzing sensor data. The context-oriented, semantic storage ensures that dynamic and ﬂexible generation of data will be more manageable and useful in the future. The integration of semantic knowledge and machine learning using multilayer neural networks results in more effi- cient and understandable data analysis. DFKI's Intelligent Networks research department is embedding modern network technologies in an intelligent connectivity platform. The use of AI procedures enables a dynamic allocation of resources, the integration of va- rious radio technologies, and continuity checks of the communication infrastructure using predictive methods. The Fraunhofer Institute for Manufacturing Engineering and Automation IPA is develo- ping a process architecture and coordination procedure for robot teams, while Fraun- hofer Institute for Industrial Engineering IAO is working on a comprehensive risk governance strategy. The latter is intended to help ensure that associated production plants in multi-organizational process structures can be proactively safeguarded in an economically viable manner. The project is based on a joint initiative of the German Federal Ministry for Economic Affairs and Energy (BMWi) and the Japanese Ministry of Internal Affairs and Communi- cation (MIC), which represented the ﬁrst German-Japanese AI project “Artiﬁcial Intel- ligence for mobile industrial communication.” The Japanese counterpart for implementation is the National Institute of Information and Communications Techno- logy (NICT). Speciﬁcally, NICT develops the adaptive and reliable means of wireless net- work communication. MORE INFORMATION Ħ https://www.dfki.de/en/web/research/ projects-and-publications/projects/ project/airport CONTACT Ü Ludger van Elst Research Department Smart Data & Knowledge Services Ö Ludger.van_Elst@dfki.de 4 +49 631 20575 1100 AIRPoRT Kick-off at the Futurium, Berlin, in November 2019. 34 News 1_2020 I © DFKI

The research demonstrator at Hannover Messe 2019 showed a hybrid manufacturing scenario for aircraft construction. Hybrid Teams for INDUSTRIE 4.0 – Research Project Hybr-iT Completed N The collaborative project Hybr-iT set out to enable more ﬂexible cooperation between people, robots, and production plants. Spon- sored by the Federal Ministry of Education and Research (BMBF), the project came to a successful close at a ceremony held at Volk- swagen headquarters in Wolfsburg on October 30, 2019. Hybr-iT met its goal by demonstrating an innovative approach to human-robot collaboration (HRC) while breaking new ground in participative work planning: team members practiced self-organi- zation with interlaced communication and effective coordination with a focus on traditional work planning. The aim was to develop and test hybrid teams in an intelligent, in- dustrial setting where humans and robots work together with soft- ware-based assistance systems. Hybr-iT researched and tested the essential components required to optimize the planning of hybrid teams – speciﬁcally, their integration and control in an actual IT and manufacturing environment. The DFKI team succeeded in developing and evaluating a mobile HRC-capable robotic system: “Mobipick” refers to a system that consists of a robotic transport platform, an industrial robot arm, and environmental sensors. The researchers studied different forms of interaction between human and robotic team players. The in- teraction included spoken language as well as direct physical in- teraction, for example, pulling or pushing the robot arm or the end effector. Indirect forms of interaction were also included via aug- mented and mixed reality. A standards-based, device-agnostic, polyglot, communication paradigm (MQ3T) simpliﬁed the devel- opment of a complex distributed system. Hybr-iT is related to the BaSys 4.0 project, which developed mid- dleware for convertible production processes. Scientists at the DFKI Cognitive Assistants Lab (COS) used the basic research results of that project in the pre-development of a proprietary BaSys4.0 im- plementation designed to orchestrate hybrid teams in INDUSTRIE 4.0. The robotic architecture was integrated into the BaSys system by the DFKI Robotics Innovation Center (RIC) for the higher-level task planning for the hybrid team. The resource-oriented architec- ture (ROA) and the development of visualization components for the BaSys dashboard were further developed by the Agents and Simulated Reality (ASR) department. The COS team was responsi- ble for integrating all signiﬁcant results of the project partners and for demonstrating the project results at trade exhibitions. A research demonstrator shown at the Hannover Messe 2019 imple- mented the project goals using a hybrid scenario from aircraft man- ufacturing: A team of two humans and six robots jointly master several assembly tasks. The robots take on what is unergonomic, repetitive, or physically stressful for people, such as setting and seal- ing rivets on the fuselage or installing cable ducts in the wing. The workers beneﬁt from a greater degree of freedom and ﬂexibility in performing their own tasks with the robots organizing themselves around them as assistants. MORE INFORMATION Ħ www.hybr-it-projekt.de CONTACT Ü Dr. Anselm Blocher I Dr. Tim Schwartz Research Department Cognitive Assistants Ö Anselm.Blocher I Tim.Schwartz@dfki.de 4 +49 681 85775 5262 I -5306 s Key project information Project partner: • Airbus Operations • Federal Institute for Occupational Safety and Health (BAuA) • Broetje Automation • DFKI (project coordinator) • EngRoTec • KUKA Roboter • The Captury • Volkswagen Project volume: approx. EUR 10 million Project term: Nov. 1, 2016 to Oct. 31, 2019 News 1_2020 I © DFKI 35

The RICAIP-Center – INDUSTRIE 4.0 and AI as an Engine for Innovation in the Czech and German Economies The aims of the RICAIP research project are to develop and test new technologies for the industrial factory of the future and to implement the targeted use of AI in industrial robocs. In mid-2019, RICAIP competed and was selected as the second-best proposal in the H2020 Teaming Phase 2 Call of the European Union. RICAIP (Research and Innovaon Center for Advanced Industrial Producon) is presently one of the largest ongoing EU projects focusing on AI and INDUSTRIE 4.0. A funding volume of approximately 50 million euros over a six-year period is approved for RICAIP for the purpose of developing new INDUSTRIE 4.0 concepts for geographically distributed producon and producon as a service. N Earlier this year, Dr. Tilman Becker became the new director of RICAIP. Before moving to the Czech Republic, Dr. Becker spent more than 25 years at DFKI researching language understanding and human-machine interaction. “The long-term goal of RICAIP is to bring large parts of production back to Europe, closer to the consumers. The new innovative man- ufacturing systems to be developed at RICAIP will be capable of ﬂexibly producing customized products with the same efficiency as currently achieved by mass production,” said Becker. “This greater ﬂexibility will be accomplished through the use of robots and AI technologies, enabling a wide variety of high-quality prod- ucts to be manufactured efficiently.” The systems developed at RICAIP will be tested and applied in prac- tical scenarios at distributed industrial testbeds in Prague, Brno, and Saarbrücken. Virtual and augmented reality will be combined with human-robot collaboration concepts: digital twins will be created for each machine to reﬂect its characteristics and communicate with other digital twins. Digital twins of entire production units will be used in the future, making it possible to simulate production processes and communications in complex environments. With the help of AI, autonomous adaptation and a signiﬁcantly faster setup time for production will be the result. RICAIP plans to apply similar principles to optimize task assignments between humans and robots. RICAIP’s other objectives include the establishment and promotion of start-ups. The technology center is also developing the ﬁrst Dr. Tilman Becker receives the appointment certificate as RICAIP Director from Prof. Vladimír Mařík, Scientific Director of the Insti- tute of Computer Science, Robotics and Cybernetics at the Czech Technical University in Prague. 36 News 1_2020 I © DFKI system demonstrator in collaboration with DFKI and the Prague- based start-up company Pocket Virtuality. The demonstrator named “Fata Morgana” (English “Mirage”) is based on Microsoft's HoloLens sensor technology and can be used to model the environment for any location. MORE INFORMATION Ħ http://ricaip.eu CONTACT Ü Dr. Tilman Becker Director RICAIP Ö Tilman.Becker@cvut.cz I info@ricaip.eu 4 +420 22435 4104 s Project information at a glance RICAIP was launched initially as a research project in 2016 as part of the German-Czech strategic cooperation in the ﬁeld of AI research. RICAIP was started by the four founding part- ner: Czech Institute of Informatics, Robotics, and Cybernetics of the Czech Technical University in Prague (CIIRC CTU), Central European Institute of Technology Brno (CEITEC), the Center for Mechatronics and Automation Technology (ZeMA), and DFKI.

XR Works – Help for Assembly Line Workers Objects such as machines and their individual parts may be in diﬀerent operang states during assembly or maintenance operaons. For inexperienced wor- kers, assembly processes that consist of several com- plex steps can be a real challenge. Delays or even errors may result, which can increase producon costs and aﬀect product quality. N DFKI’s Augmented Reality research department is developing a system that automatically detects the object states from camera images as a way to support assembly and maintenance workers. The application is based on a deep convolutional neural network that is able to accurately and efficiently as- sess both the current state and the “pose” of an ob- ject. The pose refers to the transformation between the object and the coordinate system of a camera. It consists of a translation (distance in 3D) and a ro- tation in 3D. When the pose is known, augmented reality overlays can be placed correctly over the ca- mera image and, for a single RGB image, these va- lues can be determined directly by the network. The example of a fully automatic coffee machine shows how this information can be used in an Extended AR manuals display technical instructions for servicing complex units. A fully automatic coffee machine serves as the use case for the adaptive AR assembly aid XR Works. Reality (XR) application for assembly instructions. A virtual 3D model of the next assembly step is superimposed over the current image. It is sufficient to train the network for a new assembly process using synthetic data that is automatically ge- nerated using CAD images of the objects. The system can be used to create supporting applications for production processes or for training purposes. Extended Reality (XR) applications can support people or robotic systems. Even robots can use cameras to recognize the current state and pose of an object to automatically assemble the required mechanical components. MORE INFORMATION Ħ www.dfki.de/av CONTACT Ü Jason Raphael Rambach Research Department Augmented Vision Ö Jason_Raphael.Rambach@dfki.de 4 +49 631 20575 3740 s Project information at a glance XR Works is an ongoing project in the Augmented Vision department, under the direction of Prof. Dr. Didier Stricker. In the near future, Augmented Vision will become established as a natural user interface for the home and office. As a type of AR manual, Augmented Reality (AR) systems support the as- sembly, maintenance, and operation of production plants or private use equipment. AR concepts adaptively provide speciﬁc and supportive infor- mation in situations that require an understanding of the environment as well as the user’s intentions, actions, and interactions. News 1_2020 I © DFKI 37

Demonstrator Construction Continues at SmartFactoryKL Deutsche Messe Group realized the threat of the COVID-19 pandemic this spring and canceled the Hannover Messe 2020, setting the next trade fair date in 2021. The SmartFactoryKL technology initiative based at DFKI in Kaiserslautern has exhibited its INDUSTRIE 4.0 demonstrator as a partner of the trade fair since 2014. The year 2020 was to be the year when a completely new system would be introduced in Hall 8, built according to a vision of the future with autonomous production, or “Production Level 4.” The cancellation of the Hannover Messe 2020 does not mean the end of new ideas from the masterminds at SmartFactoryKL. “The Hanover Fair is a very important date for us. Every year, we show the current state of our technical expertise there.” Prof. Dr. Martin Ruskowski, Chairman of the Executive Board of SmartFactory Kaiserslautern and head of the Innovative Factory Systems research department at DFKI The new demonstrator is a joint development effort by experts from three research institutes and eleven industrial partners. An upgrade to INDUSTRIE 4.0, known as Production Level 4, guides the functionalities and design of the demonstrator. Four innovative functions have been integrated: manufacturer-independent module ex- change, a Gaia-X link, a modular and agile system architecture, and the latest appli- cations of AI methods. The future vision of production lies in autonomous production stations (modules) that provide services to the product being manufactured. In ope- ration, the product ﬁnds its own way independently as it progresses through the as- sembly process. People still have the central role of decision-maker and controller in this integrated scenario. Machines and AI methods take over routine work or dan- gerous tasks, while people apply their skills such as ﬂexibility, creativity, or empathy. “Ours is an integrative vision of IT, machines, and people, the original concept behind INDUSTRIE 4.0. People will always be superior to technology. Machines and AI methods are created to support people and we are advancing that same concept still further.” Prof. Dr. Martin Ruskowski Photo: SmartFactoryKL 38 News 1_2020 I © DFKI Photo: SmartFactoryKL The original timeline called for the ﬁrst phase of the technical implementation of Production Level 4, to be exhibited at Hannover Messe 2020. “The de- monstrator is ready,” said Ruskowski, “now we must ﬁnd a new way to communicate it. We have several options in this respect: an event at DFKI Kaiserslau- tern could provide a suitable opportunity.” This op- tion includes planning for visual presentations on the Internet, interactive conferences, and presen- tations. “The Corona crisis forces us to be ﬂexible in many areas,” said Ruskowski. “Remote working is suddenly a fact of everyday life for office workers. In fact, we recently held a small online membership conference. The participation was higher than if colleagues had been asked to travel. The virus will force other changes, including in production me- thods. Our visionary ideas about the production of the future support this coming new reality.” DEMONSTRATOR INTERACTIV Ħ https://smartfactory.pageﬂow.io/ production-level-4 CONTACT Ü Dr. Ingo Herbst Head of Communication and Press Spokesman Technology Initiative SmartFactoryKL Ö info@smartfactory.de 4 +49 631 20575 3406

N DFKI INTERVIEW – SIMON BERGWEILER Simon Bergweiler is the team leader for HumanMachine Systems at DFKI's Innovave Factory Systems Department headed by Prof. Dr. Marn Ruskowski 9 What do you see as potential applications for your research? We believe in such topics as cognitive factories, artiﬁcial intelli- gence (AI) for human-machine interaction (HMI) in industrial manufacturing, and AI methods applied in the areas of resource efficiency and sustainability. My own focus is on providing maxi- mum support to the people at the center of factory production in the future. The deserted factory ﬂoor is not a desirable goal. This is also the thought behind the concept of Production Level 4 (PL4): the idea is to strengthen robust production systems. Achieving this requires the agility to respond to external inﬂu- ences and make the factual situation transparent to decision- makers – the people. The way to do this is with automated setup capability and data analysis of all connected systems. 9 When did your interest in Artiﬁcial Intelligence begin, and how have AI processes changed since that time? My interest in AI began 20 years ago. As I was already fascinated by the potential of AI, after my apprenticeship as an electronics technician, I decided to continue my studies by enrolling in a computer science course. While earning my diploma and later as a member of DFKI's department of Intelligent User Interfaces headed by Prof. Wolfgang Wahlster, I worked on developing an innovative concept for multi-modal Human-Machine-Interac- tion. We have since further developed the concept into an in- teractive television system (http://swoozy.de). I transferred to DFKI Kaiserslautern, where I could apply this knowledge more di- rectly to fully networked production in the sense of INDUSTRIE 4.0. Speciﬁcally, I was attracted to the subjects of modular factory units employing “Plug-and-Produce” and intelligent wor- ker assistance, which I think hold great promise. 9 What are the greatest challenges and opportunities for AI systems? At present, the expectations set for AI methods are very high. AI applications are already ﬁnding their way into new IT systems in diverse combinations. Now is the time to communicate the op- portunities of AI to the greater society. At this stage, I believe that researchers have a duty to break down prejudices. 9 What are your favorite interests outside of your work as a research scientist? In my free time, I look for balance with my family by performing DIY projects around the house, or through sports and music. 9 Do you see any parallels to your professional work? Playing an instrument in a band or orchestra is an interaction that requires each musician to complement the others with the individual parts assigned to them and their special talents. It needs coordination by a director or a drummer before the sound of the individual instruments can become a symphony. So it is with a team of researchers from various consortium partners. In essence, each contributes their knowledge to the success of the project. On the basis of their different skills as well as through close teamwork, they gain insights that none of the team mem- bers could have attained on their own. 9 What are your current work projects? I currently head the CloudiFacturing project, which aims to identify and simulate optimization potential using digitally re- corded factory data. I am also working on a messaging solution for production in the Akzent4BaSys project that is based on middleware from the earlier BaSys 4.0 project. CONTACT Ü Simon Bergweiler Research Department Innovative Factory Systems Ö Simon.Bergweiler@dfki.de 4 +49 631 20575 5070 News 1_2020 I © DFKI 39

NEWS IN BRIEF Prof. Wolfgang Wahlster, Prof. Vladimír Mařík, Scientific Director CIIRC. Photo: Roman Sejkot/ Czech Technical University Prof. Wolfgang Wahlster receives honorary doctor title at Technical University of Prague The Czech Technical University (CTU) in Prague awarded an honorary doctorate to DFKI founding director Prof. Dr. Dr. h.c. mult. Wolfgang Wahlster at the Bethlehem chapel on campus on January 21, 2020. The CTU is one of the oldest technical universities in Europe, with over 300 years of history. Wahlster joins the ranks of other honorary title holders such as Nikola Tesla, the inventor of wireless power transmission, US President and Nobel Peace Prize winner Woodrow Wilson, and the robotics pioneer Ruzena Bajcsy at UC Berkeley. Change at the top of DFKI Supervisory Board The DFKI Supervisory Board elected Dr. Gabriël Clemens as its new chairman on November 7, 2019 in Berlin. Clemens, who is a member of the executive board at the energy service provider VSE, follows Prof. Dr. h.c. Hans A. Aukes, who has been Chairman of the Supervisory Board of DFKI for thirteen years. Aukes was honored as a DFKI Fellow at a farewell ceremony on January 24, 2020 in Berlin that was attended by Saarland's Minister President Tobias Hans and State Secretary Prof. Wolf-Dieter Lukas. During Prof. Aukes' 13- year lasting term as Chairman, DFKI experi- enced strong growth. Schwarz Group to expand cooperation with DFKI To develop new technologies and applications in retail, the Schwarz Group is intensifying its current cooperation with DFKI. It will join a group of share- holders as the ﬁrst member company in the retail trade. The Schwarz Group, with its retail channels Lidl and Kauﬂand, expects the partnership with DFKI researchers to provide practical impulses in the digital transfor- mation. In the future, research ﬁndings will ﬂow even faster into innova- tions that can improve operational processes and support the stationary retail business. State AI Board meets at DFKI DFKI Kaiserslautern, at the invitation of site manager Prof. Dr. Andreas Dengel, hosted the Rhineland-Palatinate committee to network and the exchange of actors on artiﬁcial intelligence among the stakeholders in science and in- dustry in November 2019. The Board is composed of about 30 experts and is headed by State Secretary Heike Raab, Representative of the Land of Rhine- land Palatinate to the Federal Government, with responsibility for Europe, the Media and Digital Aairs. One subject discussed was the use of digital (apps) in healthcare, speciﬁcally, their quality, safety, and functional capabilities. The Digital Care Act (DVG) passed by the Bundestag in November 2019 is in- tended to let doctors prescribe digital health applications in the future. 40 News 1_2020 I © DFKI Two Best Paper Awards for Cognitive Assistants researchers Marcel Köster and Julian Groß earned the “Best Paper” Award at ICA3PP 2019 (the Inter- national Conference on Algorithms and Ar- chitectures for Parallel Processing) in Mel- bourne, Australia for their paper on “FANG: Fast and Efficient Successor-State Genera- tion for Heuristic Optimization of GPUs.” They received another “Best Paper” award for their work on “Parallel Tracking and Recon- struction of States in Heuristic Optimization Systems on GPUs” at the International Con- ference on Parallel and Distributed Comput- ing, Applications, and Technologies (PDCAT) in December 2019. Software AG appoints Prof. Markl as member of the science advisory board Volker Markl, head of DFKI's Intelligent Ana- lytics for Massive Data department in Berlin, joined the science advisory board in February 2020. Markl was nominated to the advisory board for his outstanding research and expertise in major ﬁelds relevant to Software AG, such as IoT, streaming analytics, and machine lear- ning. Chair of the panel is Prof. Wolfgang Wahlster. More information www.softwareag.com

DFKI Bremen - an active participant in the BREMEN.AI cluster DFKI is actively involved the BREMEN.AI network, which was established in 2018. The organization serves as a forum and incubator for the exchange of new ideas by oering regular events on various expert subjects. The network provides pro- ject support and promotes the recruitment and settlement of new talent and companies. The two research departments of DFKI Bremen, Cyber-Physical Systems (CPS), and the Robotics Innovation Center (RIC), have been involved in the cluster from the very beginning and have participated in sev- eral events as speakers and discussion partners. Dr. Sirko Straube, Dr. Patrick Draheim, Elmar Berghöfer, Prof. Dr. Rolf Drechsler, and Prof. Dr. Christoph Lüth have all contributed. More information https://bremen.ai Robotics Innovation Center (RIC) at Space Tech Expo Europe DFKI RIC presented its ongoing research in the ﬁelds of AI and space technologies at the Space Tech Expo Europe 2019 in Bre- men in November 2019. Visitors to the state of Bremen's exhibit experienced not only the Micro-Rover Coyote III in action, but also learned more about Bremen as an aerospace center. Wiebke Brinkmann, team leader for the “Hardware Back- bones” team and project leader for the space robotics project “PRO-ACT,” spoke about the advances in infrastructure for fu- ture space missions and was also a panel member for the ses- sion called “Tech Roadmap: Developing Tech for Future In-Space Infrastructures.” More information www.spacetechexpo.eu NEWS IN BRIEF Professor Drechsler receives “ASP-DAC Proliﬁc Author Award” In January, the “Asia and South Pacific Design Automation Conference (ASP-DAC)” presented Dr. Rolf Drechsler with the “ASP-DAC Prolific Author Award” for his numerous publications between 1995 and 2020. Professor Drechsler, with 31 publications, is the only European among the award winners. Link to the conference https://aspdac2020.github.io/aspdac20 Professor Drechsler ranked among the most inﬂuential AI experts In February 2020, the renowned Tsinghua University China re- leased the publication “AI 2000 Most Inﬂuential Scholars,” which lists the 2,000 most inﬂuential researchers in the ﬁeld of Artiﬁcial Intelligence. The listing is the result of an analysis of articles and presentations from the most important pro- fessional publications and conferences of the past ten years. Prof. Dr. Rolf Drechsler, head of DFKI's Cyber Physical Systems department, is listed under the “Circuits & Automation” (chip technology) category. HiPEAC Best Paper Award As announced by the HiPEAC Network EU ICT at the end of 2019, the paper titled “AnyDSL: a partial evaluation frame- work for programming high-performance libraries” won the Best Paper Award 2018. The paper is co-authored by Dr. Ri- chard Membarth and Prof. Philipp Slusallek, both from DFKI. The award, which is reserved for HiPEAC members, is inten- ded to encourage scientists to present their work at confe- rences where Europe may be less strongly represented. In accordance with a decision by the HiPEAC steering com- mittee, the award is only given for a complete paper that has been presented at an international conference in the ﬁeld of computing systems. More information https://www.hipeac.net News 1_2020 I © DFKI 41

Kaiserslautern Site Saarbrücken Site Bremen Site Project Office Berlin Laboratory Niedersachsen Osnabrück and Oldenburg German Research Center for Artificial Intelligence Company Profile N N N N N N N Established 1988, non-profit organization (public-private partnership) Executive Board Prof. Dr. Antonio Krüger, CEO Dr. Walter Olthoff, CFO Supervisory Board Dr.-Ing. Gabriël Clemens (VSE AG), Chair Dr. Susanne Reichrath, Representative of Saarland’s Minister President for Higher Edu- cation, Science andTechnology, Vice Chair Locations Kaiserslautern (registered office), Saarbrücken, Bremen, Berlin (project office), Niedersachsen (Laboratory) and further operating sites in St. Wendel and Trier Shareholders Accenture, Airbus Group, BMW AG, Cerence GmbH, CLAAS KGaA mbH, Daimler AG, Deut- sche Börse AG, Deutsche Messe AG, Deutsche Telekom AG, Empolis Information Management GmbH, Fraunhofer Gesellschaft e.V., Google Inc., HARTING AG & Co. KG, Intel Corporation, John Deere GmbH & Co. KG, KIBG GmbH, Microsoft Deutschland GmbH, Münchener Rückversiche- rungs-Gesellschaft Aktiengesellschaft in Mün- chen, NVIDIA GmbH, RICOH Company Ltd., Robert Bosch GmbH, ROSEN Swiss AG, SAP SE, Schwarz-Gruppe, Software AG, Technische Uni- versität Kaiserslautern, Universität Bremen, Uni- versität des Saarlandes, Volkswagen AG, VSE AG, ZF Friedrichshafen AG Equity Holding AITC Europe GmbH, Berlin GraphicsMedia.net GmbH, Kaiserslautern Ground Truth Robotics GmbH, Bremen Key Figures 2019 Annual Budget: ca. € 57,5 million Total Assets: ca. € 133 million Professional staff: 640 Graduate student staff: 440 42 News 1_2020 I © DFKI Intelligent Solutions for the Knowledge Society N The German Research Center for Artificial Intelligence (DFKI) was founded in 1988 as a non-profit public-private partnership. It has research facilities in Kaiserslautern, Saarbrucken and Bremen, a project office in Berlin, a Laboratory in Niedersachsen and branch offices in St. Wendel and Trier. In the field of innovative commercial software technology using Artificial Intelligence, DFKI is the leading research center in Germany. Based on application oriented basic research, DFKI develops product functions, prototypes and patentable solutions in the field of information and communication technology. Research and development projects are conducted in twenty research departments, eight competence cen- ters and eight living labs.Funding is received from government agencies like the European Union, the Federal Ministry of Education and Research (BMBF), the Federal Ministry for Economic Affairs and Energy (BMWi), the German Federal States and the German Research Foundation (DFG), as well as from cooperation with industrial partners. Twice a year, a com- mittee of internationally renowned experts (Scientific Advisory Board) audits the progress and results of state-funded projects. In addition, BMBF evaluates DFKI regularly. The most recent assessment was again very suc- cessfully concluded in 2016. Apart from the state governments of Rhineland-Palatinate, Saarland and Bremen,numerous renowned German and international high-tech com- panies from a wide range of industrial sectors are represented on the DFKI supervisory board. The DFKI model of a non-profit public-private partnership (ppp) is nationally and internationally considered a blueprint for corporate structure in the field of top-level research. DFKI is actively involved in numerous organizations representing and continuously advancing Germany as an excellent location for cutting- edge research and technology. Far beyond the country’s borders DFKI enjoys an excellent reputation for its academic training of young scien- tists. At present, approx. 640 highly qualified researchers, administrators and 440 graduate students from more than 65 countries are contributing to more than 250 DFKI research projects. DFKI serves as a stepping stone to leading positions in industry and successful careers as founders of spin-off companies. Over the years, more than 140 staff members have been appointed professors at universities in Germany and abroad. CONTACT Ü Reinhard Karger M.A. Corporate Spokesperson German Research Center for Artificial Intelligence GmbH Saarland Informatics Campus D3 2 66123 Saarbrücken, Germany 4 +49 681 85775 5253 3 +49 681 85775 5485 Ö uk@dfki.de Ħ www.dfki.de

Research & Development Scientific Excellence and Transfer N N N Research Departments and Research Groups Kaiserslautern Site Prof. Dr. Prof. h.c. Andreas Dengel: Smart Data & Knowledge Services Prof. Dr. Paul Lukowicz: Embedded Intelligence Prof. Dr. Martin Ruskowski: Innovative Factory Systems Prof. Dr. Hans Schotten: Intelligent Networks Prof. Dr. Didier Stricker: Augmented Vision Saarbrücken Site Prof. Dr. Josef van Genabith: Multilinguality and Language Technology Prof. Dr. Jana Koehler: Algorithmic Business and Production Prof. Dr. Antonio Krüger: Cognitive Assistants Prof. Dr. Peter Loos: Institute for Information Systems Prof. Dr. Wolfgang Maaß: Smart Service Engineering Prof. Dr. Philipp Slusallek: Agents and Simulated Reality Bremen Site Prof. Dr. Rolf Drechsler: Cyber-Physical Systems Prof. Dr. Dr. h.c. Frank Kirchner: Robotics Innovation Center (RIC) Project Office Berlin Prof. Dr. Gesche Joost: Interactive Textiles Prof. Dr. Volker Markl: Intelligent Analytics for Massive Data Prof. Dr. Sebastian Möller: Speech and Language Technology Prof. Dr. Niels Pinkwart: Educational Technology Lab Laboratory Niedersachsen Osnabrück: Prof. Dr. Joachim Hertzberg: Plan-Based Robot Control Prof. Dr. Oliver Thomas: Smart Enterprise Engineering Oldenburg: Prof. Dr. Oliver Zielinski: Marine Perception Living Labs Testing, evaluation, and demonstration of innovative technologies in comprehensive application scenarios Advanced Driver Assistance Systems Living Lab, Bremen Ambient Assisted Living Lab, Immersive Quantified Learning Lab (IQL), Innovative Retail Lab, Robotics Exploration Lab, Smart City Living Lab, SmartFactory-KL, Smart Office Space Living Lab Competence Centers Coordination of research activities in particular areas Ambient Assisted Living, Autonomous Driving, Deep Learning, Emergency Response and Recovery Mana- gement, Semantic Web, Safe and Secure Systems, Smart Agriculture Technologies, Wearable AI N N N N N N International Scientific Advisory Board Bi-annual evaluation of publically funded projects Prof. Dr. Andreas Butz, Ludwig-Maximilians- Universität, München, Chairman Leading-Edge Research DFKI is the only German institute for computer science to participate in each of the three leading-edge research clusters: Cluster of Excellence “Multimodal Computing and Interaction” funded by the German Resear Leading-Edge Cluster “Software Innovations for the Digital Enterprise” funded by BMBF European Institute of Innovation & Technology (EIT Digital & EIT Manufacturing) Research Associations Campus Automatisierung und Digitalisierung, FIWARE, Science & Innovation Alliance Kaiserslautern, and others Promoting Young Talent DFKI is a founding member and core partner of the Soft- ware Campus to promote managerial talent in the IT industry Academic Chairs More than 140 former staff members have been appointed professors at universities in Germany and abroad Spin-offs 96 spin-off companies have created more than 2.500 highly skilled jobs Service N CertLab Certified testing laboratory for IT security. For the promo- tion of credible and trustworthy Artificial Intelligence. Committees and Academies N N N DFKI is represented by its scientific directors on numerous committees and academies: Scientific and Government Committees Big Data Value Association, Brazilian Institute of Robotics (BIR), Center of Innovation Program des Japanese Ministry of Education (COI), Feldafinger Kreis, Informationstech- nische Gesellschaft (ITG), Management Board of the Inter- national Computer Science Institute in Berkeley, Münch- ner Kreis, National Institute of Informatics (NII, Tokio), Prä- sidium des Verbandes der Elektrotechnik Elektronik Infor- mationstechnik e.V. (VDE), Program Committee of the Na- tional Aeronautics and Space Research Centre DLR, Stee- ring Committee of the German Informatics Society (GI) and others Business Committees Digital Stationery Consortium, Flexible Factory Partner Alliance (FFPA), 5G Alliance for Connected Industries and Automation (5G-ACIA) and others Scientific Academies Academy of Sciences and Literature, Berlin-Brandenburg Academy of Sciences, National Academy of Science and Engineering, German National Academy of Sciences Leo- poldina, European Academy of Sciences, Royal Swedish Academy of Sciences, and others 0 2 0 2 , 1 2 l i r p A f o s A

HUMAN CENTRIC AI N INDUSTRIE 4.0, Semantic Product Memories, and Digital Twins N Smart Data – Intelligent Analytics for Massive Data N Wearable Computing and Interactive Textiles N Deep Learning, Knowledge Management, and Document Analysis N Softbots, Digital Assistants, and Chatbots N Educational Technologies N Verification and Evaluation of Safety-critical Applications N Cognitive Social Simulation N Information Extraction and Intelligent Web Retrieval N Multiagent Systems N Experience-based Learning Systems N Visual Computing and Augmented Vision N Mobile and Collaborative Robotic Systems N Multimodal User Interfaces and Autonomous Systems N Shopping Assistance and Intelligent Logistics N Safe and Secure Cognitive Systems and Intelligent Security Solutions N Smart City Technologies and Intelligent Networks N Ambient Intelligence and Assisted Living N Driver Assistance Systems and Autonomous Driving N Cyber-physical Systems N Multilingual Technologies and Language Understanding N Business Process Management and Smart Services N Affective Computing www.dfki.de | info@dfki.de Kaiserslautern Site Saarbrücken Site Bremen Site Laboratory Niedersachsen Trippstadter Straße 122 Saarland Informatics Campus D3 2 Robert-Hooke-Straße 1 Berghoffstraße 11 D-67663 Kaiserslautern Phone: +49 631 20575 0 D-66123 Saarbrücken Phone: +49 681 85775 0 D-28359 Bremen D-49090 Osnabrück Phone: +49 421 17845 0 Phone: +49 541 386050 0