Abstract
Christoph von der Malsburg obtained his Ph.D. in Heidelberg on a subject of particle physics. He worked at the Max-Planck-Institut fur biophysikalische Chemie in Gottingen from 1970 until 1987. Since 1988 he was professor of Computer Science at the University of Southern California (USC) and from 1990 he also codirected the Institute fur Neuroinformatik at the Ruhr-Universitat Bochum. Since 2006 he is Senior Fellow at the Frankfurt Institute for Advanced Studies (FIAS). In his work, he has made important contributions in the areas of pattern recognition and computational neuroscience. KI: You led the Institut fuer Neuroinformatik from 1990 until 2006 at the Ruhr-Universitat Bochum. I remember from the time I was a PhD student in your group that you were traveling between Los Angeles and Bochum back and forth twice a year. Now you are working at the Frankfurt Institute for Advanced Studies. Do you miss Bochum and Los Angeles in any way? Of course I miss them. At that time I had a large group of co-workers in both places. I had a lot of resources that I could use to develop systems. I now essentially work without any resources apart from a computer and an office. But I must say, I am enjoying both types of activity very much. KI: What are you concerned with in your current work? I think I have to collect my thoughts. I staged my life as an attempt to understand the brain and now is cash time. I have to write down my thoughts in a coherent fashion and see how far I have gotten in understanding the brain. KI: Your group was focusing in the first place on the vision problem with the aim to take the human brain as a model. What kind of knowledge about or which aspects of the brain were useful to derive vision algorithms? I think it is very important to know what the natural visual system can do and what it cannot do and, by this, to avoid putting up problems that cannot be solved at all. This is certainly one important kind of influence. The second one is that the brain is different from all of computer work in that it does not have a separate programmer. It starts with an initial condition, which has been developed by evolution, and from then on it is on its own resources. There are two ways it has to solve problems. One is by learning from examples and the other one is by some kind of self-organization or self-interaction in order to create structures that are more self-consistent. And then, of course, studying the brain has to come up with the perspective of a very simple basic data structure, which in the brain is in the form of neurons and their connections and their activity, which can express whatever is to be expressed. This is a drive towards homogeneity and simplicity. I think that is the most important lesson we can learn from the brain. KI: How would you judge the success of your group in retrospective? Well, it is the path and not the goal that is important. I think the group has achieved a lot—number one, of course, in founding careers. There are a number of professorships that arose from my institute and also a large number of students who have found good positions in industry. N. Kruger (&) The Maersk Mc-Kinney Moller Institute, Campusvej 55, 5230 Odense M, Denmark e-mail: norbert@mmmi.sdu.dk
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call