Five years of flexible and independent research - the dream of many young scientists. For junior research group leaders it can come true. Marc Toussaint is a junior research group leader at TU Berlin - his subject is machine learning.
Looking around the tidy, window-fronted office, a robot arm wrapped in foil draws the eye. "It's still brand new", explains Marc Toussaint as he removes its protective cover. "Soon it will be demonstrating how robots learn to behave intelligently." The underlying information processing routines are the subject of Toussaint's research, conducted with his "Machine Learning and Robotics" junior research group at the Technical University (TU) Berlin.
The neurocomputing specialist is one of 300 postdocs whose projects are currently funded by the German Research Foundation (DFG) in its "Emmy Noether" excellence programme. His postdoc project has already been running for two years - the funding period for a research group is usually five years.
Individual research topics for doctoral candidatesWith the flexible funding made available to him he finances three doctoral candidates from the fields of cognitive science, mathematics and computer science whom he hand-picked for his group. They work in offices next door, conduct their research embedded into his project, but they also each have their own research focus. "Group leaders handle that in very different ways", declares the young scientist. "I want my doctoral candidates to have their own research topics. What we have in common are the methods." The four-person research group from Berlin has just returned from a one-week international conference in Canada. "We showed up there resembling a small family", recounts Toussaint with a smile. "Everyone on the team had their own paper on their own subject. It was a great success for the group."
The appeal of researchNumbers and logic already fascinated him back in his school days: "Everybody always knew that I would study physics or mathematics." And that's what he did, at the University of Cologne. "Studying at university was great. What I enjoyed most was that feeling of, wow!, something new just opened up for me."
And that remains the appeal of research for him to this day: "What questions are truly open and unanswered?" The question motivated him to switch from theoretical physics to neuroinformatics after his Diplom thesis. He completed his doctorate at the Institute of Neurocomputing of the Ruhr University Bochum with "summa cum laude". Formalistic mathematical solutions like his are currently rare in this field.
Independence abroadImmediately after his doctorate, Toussaint went abroad for two years with a research grant from the DFG. "The independence was pretty important to me", he explains. He chose the University of Edinburgh, which offered him ideal conditions for his research, and expanded his current specialisms of machine learning and robotics there.
From there, he applied to the DFG for his current post. He covered the six months until his project began as a guest scientist at the robotics department of the Honda Research Institute in Offenbach. "My work there was complementary to my theoretical orientation. It was very interesting to see the robotics laboratories, and what problems arise there." Honda Research is today one of his co-operation partners in research.
"We collect ideas together"Within the Emmy Noether programme, teaching is not mandatory, but considered desirable. Toussaint will also be holding a lecture this semester. On a typical working day, he spends the morning on his own research, prepares talks and publications.
The shared lunches with his team and other colleagues, which provide lively social and professional exchange, are important to him. "In my first year I spent a lot of time teaching my doctoral candidates the basics and methods in order to create a common basis", he recalls. "Nowadays we discuss specific content-related issues and collect ideas together." Every Monday the team reads and discusses current research literature.
Good management and soft skills are important for a career in researchLiving one's life for research also requires abilities that don't come easily to every academic. "If you want to be successful in this business - and it is a business - you need many other qualities you're not prepared for and that often have little to do with research", says Toussaint. He publishes a lot and has had to develop soft skills that are also in demand on other career paths. "Research careers benefit greatly from good management and presentation skills. Or from a glowing personality."
Three or four times a year he gives talks around the world. He has twice won a "Best Paper Award" at international conferences. Nationally he sets up co-operations by regularly presenting his basic research at research institutions.
"There is no tenure track option"Toussaint would very much like to stay in Germany. "If I wanted to go abroad, I would already have gone." His aim is a professorship. But his five-year research project does not offer the prospect of being kept on at the institute afterwards. "That's the biggest problem in the German system - there is no tenure track option. The Emmy Noether Programme is intended to bridge the gap to a professorship, similar to a junior professorship. And it does, if you're only talking about the time. But it doesn't when it comes to prospects at one's own institute", he explains. "I know that there is no way I can be kept on here - no matter how good I am."
This will also be a loss to the faculty where the project is embedded practically, but not institutionally. Toussaint has set up his research group there, acquired significant funds and therefore posts with his project, and made a name for the TU Berlin in the field of machine learning and robotics. Once his project ends, little of that will remain.
Good chance with a research profileNonetheless he considers the postdoc programme to be an important milestone in his career. He summarises: "The Emmy Noether Programme gives me the opportunity to advance my own research and gain project experience. I was able to set up my own group and can expand my own research profile. Those are important criteria. I think I stand a good chance."
academics :: July 2009
12. April 2017
12. April 2017