New teaching concept: Machine learning expertise for students and industry

Professor Matthias Hein wants to teach machine learning to students from other fields as well as to industry employees.

“There are already six lectures on the topic of ‘machine learning’ at Saarland University. Therefore, it seemed logical to expand upon this focus and open it up to people unfamiliar with the field,” explains Matthias Hein, Professor of Mathematics and Computer Science at Saarland University. Hein has already won several prizes for his machine learning research, such as the German Pattern Recognition Award in 2011 and an award from the European Research Council in 2012, and he is recognized internationally as an expert in this area. In his teaching, Matthias Hein noticed that one aspect of his area is given short shrift. “It is often overlooked that the solution to a learning problem starts much earlier than in the choice of a learning method – first it has to be figured out, in interaction with an expert from the field, what the problem actually is,” explains Hein. However, this is not feasible in the small-scale exercises or mini-projects that are done in parallel with the lectures regularly offered by the Department of Computer Science at Saarland University.

Professor Hein has therefore designed a project seminar, in which three students each, over the course of 15 weeks, solve an application problem from the ground up using current machine learning techniques. This application problem is not an artificial ‘toy example’, but rather a current question coming from one of the faculties on campus, or from a company. Therefore, not only can informatics students participate in this new type of project seminar, but also students from other fields, or even employees from industry. After completing the project, every participant receives a certificate, which describes the content, procedures, and results of the project in question. “This certificate is of particular value for future job applications in the data science field,” says Hein.

In implementing this new course, he can rely on seven other Saarland University professors who are also well known in this field, including Professor Bernt Schiele, Director at the Max Planck Institute for Informatics on the university campus. Together with Daimler AG and the Technical University of Darmstadt, Schiele has already created the “Cityscapes Dataset”, the largest dataset used to train autonomous vehicles to automatically understand traffic scenes. Currently, Schiele is working further with Toyota Motor Europe to enable computers to process the visual information at lightning speed, so that autonomous vehicles can react appropriately in critical situations.

The Federal Ministry of Education and Research is now supporting this combination of expertise and dedication with about 900,000 Euro. Among other things, this sum will serve to purchase the latest hardware, which according to the researchers is essential for the application of deep learning. In the coming winter semester, the first participants from the university and industry will be able to benefit from the roughly four-month-long seminar.

Background on the Saarland Informatics Campus

The Department of Computer Science comprises the core of the Saarland Informatics Campus at Saarland University. In the immediate vicinity, seven other world-renowned research institutes carry out research on the campus. Along with the two Max Planck Institutes for Informatics and Software Systems, these are the German Research Center for Artificial Intelligence (DFKI), the Center for Bioinformatics, the Intel Visual Computing Institute, the Center for IT-Security, Privacy and Accountability (CISPA) and the Cluster of Excellence “Multimodal Computing and Interaction”.

Press photos: www.uni-saarland.de/pressefotos

Questions can be directed to:
Professor Matthias Hein
Machine Learning Group
Saarland Informatics Campus
Saarland University
Tel.: +49 681 302 57328
E-mail: hein@cs.uni-saarland.de

Editor:
Gordon Bolduan
Competence Center Computer Science Saarland
Saarland Informatics Campus
Tel.: +49 681 302-70741
E-mail: bolduan@mmci.uni-saarland.de

Notice for radio journalists: You can conduct telephone interviews in studio quality with Saarland University via radio codec (IP connection with direct dialing or through ARD starpoint 106813020001). Please send your interview requests to the press office (0681/302-3610).