# Artificial Adaptive Systems to understand world complexity. ResearchItaly interviews Massimo Buscema

**Paolo Massimo Buscema** *(pictured)* is the Director of Semeion – Research Centre of Communication Sciences in Rome and Professor at the Department of Mathematical and Statistical Sciences, University of Colorado in Denver. He has recently published a study in which he introduces a new powerful scientific paradigm to understand natural and cultural processes, which is based on two fundamental keywords: data, as representative sample of the process we need to analyse, and Artificial Adaptive Systems, as a mathematical technique able to make explicit “non-linearity” embedded in complex processes. ResearchItaly has interviewed him.

**Paolo Massimo Buscema** *(pictured)* is the Director of Semeion – Research Centre of Communication Sciences in Rome and Professor at the Department of Mathematical and Statistical Sciences, University of Colorado in Denver. He has recently published a study in which he introduces a new powerful scientific paradigm to understand natural and cultural processes, which is based on two fundamental keywords: data, as representative sample of the process we need to analyse, and Artificial Adaptive Systems, as a mathematical technique able to make explicit “non-linearity” embedded in complex processes. ResearchItaly has interviewed him.

**Professor Buscema, first of all, what are complex systems?**

A system is a compact region of space–time, divided into components, whose local and parallel interactions determine the functioning of the system itself. A system in which the structure and the status of its members dynamically change, while maintaining its cohesion space–time, is called a complex system. A human cell that becomes a human organism is a complex system, an aircraft, with all its thousands of small components, is not. A cell generates new information during its evolution, while an aircraft does not generate any new information during its construction, every possible trajectory can be defined *a priori*, therefore it is not a complex system, it is only a very complicated system. As it is often said, in a complex system, the global behaviour of the system is not inferred from the simple sum of the behaviour of all its components, its components actually undergo structural changes over time, since a complex system is adaptive. This is why to understand its functioning non-linear analysis techniques are necessary.

**What are Artificial Adaptive Systems and what are they used for?**

The Artificial Adaptive Systems new mathematical algorithms understand complex systems: they are the mathematical expression of the natural computational method that does not impose linear assumptions to the data. We can say that the mathematics that is needed to understand natural phenomena works on data using a Socratic style, i.e. it works in a way that the overall behaviour of the phenomenon emerges spontaneously by the interaction of its components, represented by data and equations, in the so-called bottom-up system. Conversely, in natural adaptive systems, there is a sort of hidden information that reveals the secret plan that the system is going to pursue. Artificial adaptive systems look for traces of this hidden plan.

**In the light of this, what is the protocol that can support current scientific research?**

The protocol for scientific research is simple and consolidated over time. It can only focus on three keywords: Research, Experiment and Simulation. “Research” in this context means the entire investigative-experimental operation, from the identification of the targeted problem up to the interpretation of the results; by “Experiment,” we mean the design, execution and completion of data processing, from the initial treatment of available data up to the interpretation of the local results; by “Computer Simulation,” we mean the process that starts from the application of the different artificial adaptive algorithms to the data, up to the generation of results.

**Can you make an example of how an Artificial Adaptive System can be applied to a concrete problem?**

Now, let us imagine a certain area where a series of robberies has occurred. We are observing a series of events that have occurred in a given physical space. The questions raised are: are those events connected? Are they the work of the same persons? Are the robbers strangers or do they live in that area? Do the robbers live in the same neighbourhood or elsewhere? Where do they live? If the only data available is the total number of the robberies and the place where they have occurred, and we define the place with longitude and latitude values, how can we make a reasonable estimation about the place where the perpetrators might live? Translated into our own language: we have a distribution of events, generated by the same process, into a two-dimensional space that has a great deal of hidden information: one of these is the possible location from which all of the events originated, i.e. the “event zero” of the space-time sequence.

**Have you developed any specific algorithms to study this type of phenomena?**

Yes, at the Semeion Research Centre we have recently developed the Topological Weighted Centroid (TWC) that is useful to solve this type of puzzles. We used it to disrupt a gang of robbers in Denver, Colorado, in 2011, thus supporting the local police to bring the perpetrators to justice. We used it for the analyses and simulations of the geodynamics of Ebola virus outbreak in 2014, in collaboration with the University of Colorado and the Rocky Mountain Poison Centre. We also used it to study the spreading of the outbreak of Dengue fever in Brazil, the *escherichia coli *mutation outbreak in Germany in 2011, the Listeriosis epidemic in the United States and – back to Italy – the famous case of Unabomber and the toxic burnings in the provinces of Naples and Caserta at the beginning of 2013.

**Based on your experience, what is your advice to young researchers?**

Carry out as many experiments as you can and, if you fail, it is even better because you will have the opportunity to grow: a failed experiment is the necessary condition for substantial change and innovation. Pay attention to any actual or potential imperfections in concepts, techniques, models that are considered as dogmas by tradition. Learn to observe each process from the same perspective that nature used when generating it. Pay attention to the relationship between objects, because relationships come before objects. Learn to listen to the weak bonds between things because weak bonds can often explain stronger bonds. Use mathematics to turn a simple scribble into a possible pattern since nature, whenever possible, avoids noise…