Foundations for trustworthy artificial intelligence

Machine learning and the applications of artificial intelligence are of growing importance in science, economy and society. Artificial intelligence (AI) refers to technologies that enable computers to help humans with tasks that can only be solved by intelligence. Statistical and data-driven methods known as machine learning can generate valuable results from very large, complex or heterogeneous data sets.

Governments, corporations and universities the world over are implementing strategies to address AI’s growing economic and social impact, with the US and China investing particularly heavily. One initiative that connects AI researchers across Europe is being inaugurated today at a external pagevirtual ceremonycall_made: the European Laboratory for Learning and Intelligent Systems (ELLIS). The ELLIS network spans 14 European countries and several world-class institutions, among them the Max Planck Institute for Intelligent Systems, University of Cambridge, University of Oxford, Technion Israel Institute of Technology, Sorbonne Center for Artificial Intelligence, EPF Lausanne, and ETH Zurich.

30 ELLIS units join forces

Initiated in December 2018, ELLIS today comprises 30 research units. ETH Zurich was chosen for one of the first 17 ELLIS Units in 2019. “With their combined strengths, the units will contribute to enabling Europe to compete with the world’s AI hotspots, particularly the US and China,” said Bernhard Schölkopf, co-founder of ELLIS and Co-Director of the Max Planck ETH Center for Learning Systems, at the ELLIS virtual inauguration ceremony today. “Together, the units are creating new opportunities for collaboration with scientists across Europe, and laying a strong foundation for developing AI that is in line with the values of open European societies.”

In a video message at the inauguration, ETH Zurich President Joël Mesot stressed the importance of joining forces in order that Europe advance independently in AI alongside the USA and China. “ELLIS will give us the chance not only to make our economies more dynamic and competitive, but also to develop an AI that meets ethical requirements,” said Mesot. “ETH Zurich is keen to contribute its scientific expertise to this initiative and to help enable technologies that serve society in an integrated way.”

The central goal of ELLIS is to foster research excellence and technology transfer in machine learning and related fields, and to train the next generation of talent through Europe-wide network activities. In total, the ELLIS units have committed funding of some 300 million euros for an initial period of five years.

ETH emphasises top talent and collaboration

The ETH Zurich ELLIS Unit is an integral part of the new ETH AI Center, which will be opened in a virtual ceremony on 20 October 2020. On board are more than 25 professors from six ETH departments: Mathematics, Computer Science, Information Technology and Electrical Engineering, Mechanical and Process Engineering, Biosystems Science and Engineering, and Civil, Environmental and Geomatic Engineering. One third are female researchers.

The ETH AI Center connects the AI researchers within ETH and combines the fundamentals of AI theory and methodologies with expertise from the various disciplines.

“Our ETH AI Center and ELLIS Unit are dedicated to interdisciplinary collaboration in AI research. All our facilities are open to partners and guests from science and industry,” says Andreas Krause, who chairs the ETH ELLIS Unit and the ETH AI Center. “We are looking to provide an exciting, integrative and stimulating environment for top talent, to support start-ups, and to foster dialogue with policymakers and the general public.”

Focus on trustworthiness, health and sustainability

Research activities at ETH Zurich range from the theoretical foundations of machine learning and the principles of a reliable and trustworthy AI to the study of how AI systems interact with humans and assist them. Interdisciplinary applications in domains such as health, sustainability and environment, and robotics are also a focus, as well as the study of the ethical, societal and policy implications of AI. “When it comes to societal impact and trust in AI, reliability and transparency are essential,” says Krause.

Within the ELLIS network, research and exchange are organised in 11 ELLIS programmes and in the “ELLIS against COVID-19” initiative. ETH researchers are co-directors in two ELLIS programmes: Andreas Krause co-directs the external pageinteractive learningcall_made programme and Gunnar Rätsch, Professor of Biomedical Informatics, the external pageELLIS Healthcall_made programme. In addition, Luc van Gool, Professor of Computer Vision, and Konrad Schindler, Professor of Photogrammetry and Remote Sensing, are Fellows in the external pageComputer Visioncall_made and external pageEarth and Climate Sciencescall_made programmes. Other ELLIS Fellows are Joachim Buhmann, Thomas Hofmann, Karsten Borgwardt and Davide Scaramuzza.

“The ELLIS Health programme means I can help build an active research community in the field of Machine Learning for Health in Europe and Switzerland,” says Gunnar Rätsch: “This will create a fertile climate to drive research and innovation in machine learning and AI in Europe to the next level.”

For Julia Vogt too, a big asset of ELLIS is the exchange. The ETH Professor in Medical Data Science is developing machine learning techniques that assist doctors in treating diseases more effectively. “The latest results and AI research issues will be openly discussed in the workshops and made quickly accessible,” she says. “Through ELLIS, I am discovering new synergies, and these have already led to outstanding collaborations in Europe.”