PhD in the Spotlight: Rick Quax
On 14 March, Rick Quax (b. 1985) will receive his doctorate degree from the Informatics Institute (IvI), where he has developed a new theory to predict how complex systems function. He has discovered that in the short term, individual elements that interact intensely with other elements have little effect on the system as a whole.
What was your most important finding?
‘I have developed a new theory that improves our understanding of complex systems. There are two types of systems: regular and complex. It is possible to predict the behaviour of regular systems as long as you know the behaviour of their individual components. This is not possible with complex systems, in which individual behaviours are so convoluted that they cannot be traced. The human brain is one example. Even if you know how neurons behave at an individual level, we still do not know how the brain functions as a whole. My theory predicts that the more interactions a neuron has with other neurons, the less influence that neuron will have in the short term. Although it sounds contradictory, we think it is a widespread phenomenon in nature.’
How did you make this discovery?
‘In two ways. I first demonstrated the theory mathematically, by writing it out and applying algebra to it. Later I ran simulations using the SARA supercomputer, where I had it run continuous parallel calculations for hours. Applied to a model of complex systems, both methods demonstrated that the more a component interacts with others, the lower its influence on the system as a whole. I then compared our prediction with a range of complex systems that have been the subject of comparable experimental research, such as the network of protein interactions within a cell, or word-of-mouth advertising in social networks.’
And were your results the same?
‘Yes. The experiments showed an inverse relationship between the number of interactions of a gene, neuron or person and its influence, with no explanation. Our theory therefore presents a plausible explanation for this phenomenon. It doesn’t work for all complex systems, however: epidemiological models are one exception. The more contact an infected person has with others, the more people become infected, thereby increasing the individual’s influence on the entire system.’
Will you be conducting further studies?
‘I will be continuing this research in various ways as a post-graduate under my supervisor, Peter Sloot. In conjunction with our fellow researchers at the Swammerdam Institute for Life Sciences (SILS), we plan to apply the theory to neural networks and to conduct more experiments. At SILS, research is being conducted on neuronal changes associated with epilepsy. After a seizure, the network structure among neurons changes, resulting in more neurons with few connections, and fewer neurons with many connections. The extent of this effect on the brain is still unknown, but we believe that we can quantify it using my theory. I also plan to look at new metrics for information transfer within complex systems. I always start on the theoretical side, and once I have a new theory I seek to validate it. That’s what I enjoyed most about my doctoral research – that instead of being purely theoretical, it also had a substantial practical aspect.’
Author: Carin Röst