A boost for artificial intelligence

How can a computer learn to generalize the knowledge contained in a huge dataset and to find solutions for complex problems on its own? This question is the focus of two new assistant professorships at the Faculty of Engineering at the University of Freiburg: Robot Learning and Representation Learning



How can a computer learn to generalize the knowledge contained in a huge dataset and find solutions for complex problems on its own? This question is the focus of two new assistant professorships. Photo: Jürgen Gocke

They are the result of a successful application in the program "Artificial Intelligence Baden-Württemberg" administered by the state Ministry of Science, Research and the Arts. The professorships will join the research underway at the Faculty of Engineering. The state of Baden-Württemberg is sponsoring the professorships over the next four years with a total amount of 1.2 million euros. The positions are to be filled by early-career researchers by May 2019.

“Whether it is data evaluation, autonomous driving or smart homes, Industry 4.0 or in academia: It is of vital importance for our society, academia and economy to research the development artificial intelligence methods and their applications,” says Science minister Theresia Bauer. “The two assistant professorships will be working in the new Intelligent Machine-Brain Interfacing Technology research building and will contribute to further reinforcement of the University of Freiburg’s position as one of the leading locations in Germany in the field of artificial intelligence,” says Professor Dr. Gunther Neuhaus, Vice-President for Research and Vice-Rector at the University of Freiburg.

The greatest progress in artificial intelligence (AI) is currently in the techniques of machine learning. That is when a program learns by itself from the data given to it, and is subsequently able to apply this knowledge more generally and specifically to other problems. In particular, robotics has great potential to improve tasks such as perception, prediction and navigation - for example in robots which can reliably recognize objects or to reach out and touch them. The Robot Learning junior professorship has the aim of developing methods enabling robots to conduct their tasks more effectively and robustly and to adapt more flexibly to their environment.

A further trend in artificial intelligence is deep learning, a technology based on artificial neural networks. With the help of the very large data sets and computing power available today, these networks learn, via many layers of processing, to represent data more abstractly. Early layers extract simple patterns from raw data - such as the recognition of edges in image processing. Later processing layers combine these patterns with ever more complex and ever more abstract representations, so that a computer can ultimately make robust classifications - for example, in the context of self-driving cars, it learns to distinguish humans from lampposts or to detect traffic lights. The Representation Learning assistant professorship seeks to improve and further develop the basic methods of deep learning.

Artificial intelligence is an established field of research at the University of Freiburg’s Faculty of Engineering. Seven professors and their groups are currently conducting research into various aspects of artificial intelligence, covering relevant topics such as robotics, algorithms and data structures, machine learning, computer vision, neurorobotics and the basics of artificial intelligence. The Intelligent Machine-Brain Interfacing Technology (IMBIT) building, which is to open its doors in early 2020, will provide researchers with outstanding research infrastructure and will support the exchange of ideas.

Baden-Württemberg Ministry of Science, Research and the Arts press release

Professor Dr. Hannah Bast
Faculty of Engineering
University of Freiburg
Phone: 0761/203-8163
E-Mail: bast(at)informatik.uni-freiburg.de

Natascha Thoma-Widmann
PR/Marketing, Faculty of Engineering/BrainLinks-BrainTools
Phone: 0761/203-8056
E-Mail: thoma-widmann@tf.uni-freiburg.de


Benutzerspezifische Werkzeuge