Emmy Noether success for ZESS member Jovita Lukasik

A major success for AI research in Siegen: ZESS member Jun-Prof Dr Jovita Lukasik has secured funding under the prestigious Emmy Noether Programme run by the German Research Foundation (DFG).

Up to now, the development of a neural network has involved two separate steps: first, the structure of the network is defined. In the next step, the network is trained using data. The idea behind Jovita’s project “UniCoDe” is to automate the merging of these two steps: Instead of being laboriously designed by humans and then trained, the networks are intended to build themselves in future whilst simultaneously optimising themselves and adjusting their parameters to achieve the best possible results.

“There has been a lot of hype for some time now surrounding AI systems like ChatGPT, which are based on deep networks and huge training datasets. I’m not interested in building ever-larger networks. I’m more interested in how systems can be designed more efficiently. That’s what I want to focus on with the Emmy Noether Group,” says Jovita who is also pursuing a second goal: “I want to encourage young women to conduct research in computer science and AI. It’s an exciting field with many opportunities.”

Jovita took a position as Jun-Prof for Visual Computing with University of Siegen only six months ago and has since also been a ZESS member. Succeeding with her application for the Emmy-Noether-Programme, which is one of Germany’s most prestigious programmes for early-career researchers – establishes the first Emmy Noether group ever at the University of Siegen. The project is being funded with around 1.6 million euros, a total of 178 project proposals for new Emmy Noether Groups in the field of artificial intelligence were submitted to the DFG in this round of applications. Only 15 of these were selected for funding.