Complexity is a subtle concept that comes in a wide variety of flavours. It is associated with systems composed of large numbers of components interacting through numerous positive and negative feedback loops, thus giving rise to global behaviour that is notoriously hard to predict. This type of complexity pervades every level of our world, from multicellular organisms to urban infrastructures, and from socio-economic networks to climate change. It is an intriguing fact that these systems often show striking parallels in the ways in which they behave or evolve, despite their differences in scale and organizational detail. Examples of this include the emergence of new and hierarchically ordered levels of internal organization, the occurrence of global transitions triggered near tipping points, and even the vulnerability to catastrophic collapse due to cascading failures propagating throughout the system.
However, complexity also appears in a different guise at an even more fundamental level, where it is closely related to pattern extraction, compression and randomness. As such, it is instrumental in opening up new methodologies for information extraction from data sequences, suggesting novel principles for machine learning, and providing insights into the limits of computability. At CWI, we are interested in all these aspects of complexity. Related questions that we face include: how can we build computational models that describe all relevant interactions across a wide range of spatio-temporal scales? How do we efficiently handle the massive amounts of computation and data that are required for reliable simulations? And how can we design local interactions in complex systems so that they result in global behaviour that is controllable, efficient and robust?
Investigating complexity at an abstract, rather than a specific, level means that our findings are applicable across a wide range of applications and can therefore be used by a host of other disciplines.