Abstract. Some complex systems are naturally composed of well-defined building blocks from a heterogeneous shared set. A toy example of such a system is a lego set. The different target architectures, a farm, a hospital, or the millennium falcon translate into different statistical patterns of the bricks. Analysis of many Lego sets can in principle reveal common and specific features of different objects that can be built with Lego. A more intricate example of such a system is a genome, seen as a set of gene families with common evolutionary and functional features. In this case, we only have partial information about the architectural constraints and about the structures that are under natural selection. We propose that these "component systems" can be understood with a common theoretical framework and common data analysis tools. Direct analysis of empirical data on component systems in several contexts (genomics, ecology, quantitative linguistics and technological systems) supports this view. In this seminar, I will discuss the main trends and the current state of our modeling efforts.
This talk will stream live from our youtube page: https://www.youtube.com/user/santafeinst