Making Machine Learning More Democratic

Direktzugriff

Artikelaktionen

Professor of Computer Science Frank Hutter wants to democratize machine learning and help Germany and Europe stay competitive on the global market. Photo: Patrick Seeger

Freiburg, Jan 13, 2020

Machine learning (ML), which is a form of artificial intelligence in which systems can make their own predictions based on data, has become a key technology in manufacturing. However, most small and medium-sized companies are rarely able to benefit from ML because they can’t afford the necessary expertise. Prof. Dr. Frank Hutter and his team at the Department of Computer Science of the University of Freiburg have been awarded a Proof of Concept Grant for 150,000 euros from the European Research Council (ERC), which they plan to use for further developing their automated machine-learning (AutoML) prototype. Through AutoML, a machine can automatically determine which machine learning processes work with what settings. The researchers want to use their prototype to make it easier for non-experts to use and afford ML methods.

The ERC awards Proof of Concept Grants to previous ERC grant recipients who want to evaluate the commercial potential of their work from a current or a past project that they believe will help Europe to get ahead of global competitors. Hutter and his team of researchers received a Starting Grant of 1.5 million euros from the ERC in 2017. With the help of this funding, the scientists have developed a research prototype that has won several competitions. Their goal is now to continue to develop this into a professional prototype, using the results of the new project as a basis to create a sustainable business model. For this purpose, they will be investing in market research, testing their system in a manufacturing environment, and checking the prototype’s potential for a market launch. “AutoML lowers costs by saving computer resources and human hours at the same time. Small and medium-sized companies could especially benefit from this, because they aren’t able to afford a big team of machine learning experts,” says Frank Hutter, adding: “AutoML lets us make machine learning more democratic and helps us to keep Germany and Europe globally competitive.”

Press release about Frank Hutter’s ERC project

Website of Hutter’s Automated Machine Learning (AutoML) project

The European Research Council (ERC)

Contact:
Prof. Dr. Frank Hutter
Department of Computer Science
University of Freiburg
Phone: +49 (0)761 / 203-67740
E-mail: fh(at)informatik.uni-freiburg.de

Fußzeile

Benutzerspezifische Werkzeuge