New centre for AI research

ETH Zurich is opening a new research centre for artificial intelligence. A core team of initially 29 professorships, a new executive director and a Fellowship programme aim to promote interdisciplinary research into this key technology.
The newly opened ETH AI Center intensifies the dialogue with business, politics and society about an innovative and trust-​building advancement of artificial intelligence. (Photograph: ETH Zurich / ETH AI Center)

The following example illustrates why AI is an important technology for society as a whole. Jaundice is one of the most common diseases in newborn babies, with about 60 percent showing symptoms. In approximately 10 percent of the babies affected, the disease becomes critical and can have lifelong consequences if not treated. Diagnosis of the disease is straightforward and it can be treated with light therapy. However, with increasingly shorter hospital stays after birth, the symptoms may not appear until the baby is already at home.

With the aid of a new AI model, researchers at ETH Zurich have succeeded in developing a new app that can predict the disease in babies 48 hours before the first symptoms appear with just four indicators. “This app was developed in close collaboration between AI researchers and medical professionals,” explains Julia Vogt, a professor of medical data science. Interdisciplinary cooperation can be a major challenge; for example, due to different ways of thinking and a lack of awareness of the specialist knowledge of the other parties. The opening of the ETH AI Center – an ETH-wide centre for AI research – provides ETH Zurich with the space and resources to intensify AI research in terms of foundations, applications and implications.

“The ETH AI Center will intensify interdisciplinary dialogue with business, politics and society through the innovative and trust-building advancement of artificial intelligence,” says Detlef Günther, Vice President for Research at ETH Zurich. The centre will also become an incubator for AI start-ups and facilitate joint research projects with international experts in an Open Lab. In addition, it will become part of the European AI research network ELLIS.

New space, new leadership

The centre has already moved into its premises in the university district, and new office space and laboratories will be available in the Andreasturm in Oerlikon at a later date. PD Dr Alexander Ilic, the centre's new Executive Director, is an entrepreneur and most recently headed Magic Leap Switzerland. “I am delighted that Alexander Ilic, as a renowned AI expert, agreed to take on this role. He is a successful AI entrepreneur, knows the academic world and has a strong network in the AI scene,” says Andreas Krause, professor at the Institute for Machine Learning and Chairman of the ETH AI Center.

Core team with around 29 professorships

The new research centre is made up of core members, associate members and AI fellows. The core members already comprise 29 professors from seven departments who specialise in fundamental AI topics such as machine learning, perception and natural language understanding. These members will also supervise the new fellows of the ETH AI Center, talented young AI scientists who have been recruited from around the world and awarded a scholarship. “This will enable us to train experts who will be able to implement interdisciplinary AI projects,” says Günther. The ETH AI Center will also maintain contact with associate members from all ETH departments, other institutions and the private sector in order to implement joint research projects.

Artificial intelligence based on ethical principles

Researchers at the ETH AI Center aim to develop tools that will help people in various areas to solve complex tasks efficiently. It is essential that AI models work reliably and that their results are robust, explainable and fair. “Reliability and interpretability are essential to AI research,” explains Krause. “These aspects relate to highly relevant questions about the social impact and ethics of AI.” Reliable AI solutions help to increase the acceptance of this technology in the public domain, particularly in research areas such as health, mobility, architecture, energy and climate, in which ETH Zurich has an excellent position. “Widespread acceptance is a key competitive advantage,” emphasises Krause.