It is hard to envisage scientific technical research today without the use of more or less sophisticated computer models. Even though the open questions in certain research areas are very different, the developers of such programmes are often confronted with similar methodological problems. It is often a question of simulating complex systems or representing processes in a model that take place on diverse time and space scales, and which are still only partially understood.

Making solutions available

The ETH Zurich wants to make the body of existing knowledge in many departments available to interested parties from other disciplines in order to diffuse successful solutions to those areas. CoLab (1) was founded in May 2002 with this goal in mind and researchers from a number of departments make up the participants. The name is short for "Computational Laboratory" but also a (somewhat dodgy) play on the word CoLab-oration. Last October workers moved into new office rooms in the Hirschengraben near Central in Zurich where they are developing methods to solve complex scientific problems.

Software more important than hardware

"Everyone knows that hardware has become more compact in recent years and that this has made even more efficient computer models possible. What is not so well known is that the algorithms, which lie at the heart of the models, have also developed exponentially" says Petros Koumoutsakos, Professor of Computational Sciences at the ETH Zurich and he adds, "improvements in the software have done even more to increase the efficiency of the models than hardware developments." For CoLab's director of the steering committee it is clear: "In order to improve models even more, we don't just need quicker computing power and more capacity but cleverer calculating methods."

Great interest from young scientists

CoLab is not an isolated institution but the core of the programme "Computational Science and Engineering" (CSE). In its turn, this is one of four projects that the direction of the ETH has initiated and given special status to under the title of "Strategic Success Positions" (SEP). Between ten and fifteen postgraduate positions are planned for CoLab.

According to Sabine Attinger, scientific co-ordinator at CoLab, "Young scientists, in particular, show a great interest in the new institution." Around 40 candidates were selected from 400-odd original applications and invited to interviews. In the meantime, 11 postgraduate positions have been filled.

Division of labour

In order for CoLab to become the networked platform that its initiators want it to be, the postgraduates have to work closely with research teams from the ETH. This prevents them from becoming isolated in their research. Attinger hopes that the ambitious goal can be achieved: CoLab is a place where an intensive exchange between scientists from diverse specialist backgrounds takes place and, at the same time, close co-operation develops and is upheld with outside institutions.

Model simulations play an important part in many of today's scientific and technical disciplines.

Focusing on subject and method

For the main part, research projects at CoLab come from the areas of cell biology, bioinformatics and materials science. Further on the list are also projects associated with CoLab from other disciplines, such as earthquake research, transport technique and climate research. Two main problems areas are being addressed, explains Koumoutsakos, with "multiscaling" and "machine learning".

Multiscaling addresses the question of how to describe processes that run on different scales–and therefore have different physical solutions–and to combine them in one model. Machine learning has to do with the complex modelling of data and rule based complex systems that cannot be described using "simple" mathematical-physical conformities. Biological systems, for example, belong in this category. In order to model them, one has to depend on observation and data. The aim is to work out rules for the behaviour of such a system on the basis of such data in order to better understand the system.

A veritable challenge

In addition to concrete research the heads of CoLab seek to foster the exchange among scientists by the means of numerous complementary activities. Renowned scientists are invited to visit CoLab under the "Visiting Faculty Programme" and to stay for a certain period, working with the permanent staff. This provides interesting contacts for the postgraduates. Next summer CoLab, together with the CSCS in Manno and the Università di Svizzera Italiana in Tessin, will host an international summer workshop lasting several weeks on "Multiscale Modelling and Simulation". CoLab intends to publish the results of this event in book form. Finally, CoLab wishes to build up an intensive dialogue with biologists. The aim is to create a mutually fruitful learning situation for both computational scientists and biologists by making close contact feasible. This seems to be a somewhat difficult undertaking. Or, as Koumoutsakos diplomatically puts it, "Finding a common language is the veritable challenge."