membership declined to about 40 members and
most universities did not have an AI chair or offer
AI courses in their curricula. One notable exception was École Polytechnique Fédérale de Lausanne
(EPFL), where Boi Faltings and his team made significant contributions to the area of constraint
optimization. Other research groups would also
continue to work in areas such as robotics, computer linguistics, or search algorithms, but would
often not see themselves as part of the worldwide
AI community that gathers at conferences such as
AAAI, IJCAI, or ECAI.
The relative quietness of AI in Switzerland did
not prevent companies looking at AI technology as
a potential key driver for innovation. One example of this was Schindler, a worldwide maker of elevators. Schindler became interested in AI planning
after the first AI planning competition in 1998 and
soon after started to develop an AI-based control
algorithm for its products based on destination
2 Today, these algorithms are the most
sophisticated in the elevator market and made the
company’s destination-control products a huge
commercial and award-winning success. Many
other, smaller, technology-savvy companies and
startups enjoyed similar successes, all of which
have led to a growing need for AI specialists in
Current Lines of Research
With the recent AI-based technology successes,
interest in AI in Switzerland has resurged. In the
fall of 2012, the Swiss Informatics Society put
SGAICO under direct guidance of the SI president
to prevent the closure of this special interest group.
A diverse team of researchers was asked to help
revive the group and a small amount of money was
granted to restart it. The organizing team included
researchers from the University of Basel, EPFL Lausanne, and three universities of applied sciences in
Zurich, Lucerne, and Western Switzerland (
HES-SO), who brought in numerous academic and
industrial contacts. The team has launched several successful events and continues to bring energy
to the AI and cognitive science community.
SGAICO puts specific emphasis on joining contacts
between universities, applied research, and industrial applications and on reintegrating the AI and
cognitive science communities within the country,
where of course much remains to be done.
In the following paragraphs, we provide selected
pointers to groups in Switzerland that work in specific subfields of AI.
and Constraint Reasoning
The areas of planning, scheduling, and constraint
solving are currently being studied at the Universi-
ty of Basel in the group led by Malte Helmert. A
team at Lucerne University of Applied Sciences has
a similar technology focus with a strong emphasis
on embedding these technologies into product
innovations for Swiss companies. Current projects
in Lucerne include control algorithms for major
Swiss players in the machine industry, applications
in medical diagnosis, forensics, and social welfare.
Machine Learning and Data Science
The Swiss IDSIA lab is a pioneer and major proponent of deep learning. Jürgen Schmidhuber’s team
at IDSIA recently won several prestigious pattern-recognition competitions using deep neural architectures.
ETH Zurich has appointed several chairs working in machine learning and, for example, organizes the Zurich Machine Learning and Data Science
Meetup, which easily attracts around 100 attendees for an evening of three technical lectures.
One possible explanation for the recent and very
strong interest in machine learning might be the
trend toward big data that has been widely recognized among industry and academic representatives alike. Much interdisciplinary work is done
under the headline of data science. For example,
the Datalab at the Zurich University of Applied Sciences is one of the first interdisciplinary groups in
Europe devoted to the creation of data products for
and with the industry. Recent results include medical applications, data-driven condition monitoring approaches, and specialized information
Information Retrieval, Natural Language
Processing, and Computer Linguistics
A lively and internationally well-connected information retrieval community exists with notable
groups at the University of Geneva, the Universities of Applied Sciences of Western Switzerland
(HES-SO) and Zurich, the University of Neuchatel,
IDIAP (an Institute affiliated with EPFL), and the
Università della Svizzera italiana (USI). As an example, we mention the CLEF and ImageCLEF yearly
5 in which Henning
Müller and Martin Braschler are engaged.
Work in natural language processing and computer linguistics is also strong, in particular in
Zurich and the IDIAP institute. An interesting
application of natural language processing is, for
example, the paperless parliament in the Canton
of Valais, where all protocols are generated automatically based on algorithms developed at IDIAP.
Multiagent and |
Cognitive Systems, Robotics
Another subfield that has attracted both theoretical and practical research is multiagent systems.
There are several Swiss research teams, such as