The group was started in April 2012 und combines research in the area of "Complex network analysis",
"Algorithmic graph theory",
and "Computer assisted cognitive science". Due to the interdisciplinary education of Professor Zweig and the multi-disciplinary
group consisting of mathematicians, physicists, and computer scientists, most projects are based on an interesting data set from
some of our cooperation partners, be it from cancer research, ecology, economy, psychology, or law. We then represent the data
as a complex network and extract the most significant patterns in it ("Complex network analysis").
In most cases, we will develop new algorithms for this task,
based on the needs of our cooperation partners ("Algorithmic graph theory"). A special focus is on developing and supporting
computer-supported psychological experiments (
"Computer assisted cognitive science").
We are always interested in new cooperations and look forward to your request.
Complex network analysis:
At the end of the 1990s, complex networks came into the focus of statistical physicists. A complex network represents the interactions
of subjects and objects in a complex system. Typical examples of complex systems are social systems and any type of biological cells.
A complex network normally represents interactions between only one type of entity of the systems, e.g., the communication between
humans in a given social group or the protein-protein interactions of a cell. Based on its network representation, we use methods
from social and complex network analysis to find significant subpatterns in it, to look for central nodes in the network or
to identify groups of similar subjects and objects.
Examples for such projects by us are the identification of cell growth inhibiting microRNAs in a breast cancer type with especially high lethality [Uhlmann et al., 2012], the insight that social network platforms like facebook can induce quite a lot of information about non-members [Horvat et al., 2012], or the question of how the similarity of two products can be deduced from the buying behavior of products [Zweig, 2010].
Algorithmic graph theory In some cases, the request of a cooperation partner leads to a new and interesting theoretical problem in graph theory. Examples for this are our finding that finding all cliques in a max-tolerance graphs is easy - based on the question of how to cluster the aligned sequences of a BLAST result [Kaufmann et al., 2006][Lehmann et al., 2006]. Another examples is our work on the conditions under when classical association rules cannot be used [Zweig 2010] and on one-mode projections of bipartite graphs [Kaufmann and Zweig, 2011]. Next to this we have worked on cycle bases [Kavitha et al., 2009] and on a drawing model for streamed graphs [Binucci et al., 2010].
Computer assisted cognitive science A new focus of the group is on the question of how we can assist psychological experiments by providing efficient algorithms to analyze thousands of results in a short time. Most classic psychological experiments are tun under lab conditions in which participants take part in a standardized environenment and way. This way of experimenting is reliable but time consuming and costly; it also only allows small numbers of participants (less than 100, in most cases). In many cases, the experiment itself is already run on a computer and in principle it is easy to just put it on a website and let many participants try it. However, many of the participants might start an experiment and then go get a coffee or be called by someone. To make the analysis reliable it is necessary to automatically detect these deviant patterns. As the numbers are much higher than usual, it is also much more necessary to have an automated analysis of the results which can be more or less difficult. We are especially interested in the problem solving capabilities of humans in complex situations; we test it by letting them play games like RushHour. Our first project in that direction was the analysis of how humans find their way in a highly abstract network [Iyengar et al., 2012]
|Superordinated page: Graph Theory & Complex Network Analysis Group|
|Homepage of the Graph Theory & Complex Network Analysis Group [more ...]|
|Feedback for this Page|
|Email to the person in charge of this page (firstname.lastname@example.org) [more ...]|