They contribute to meeting 21
century challenges, notably in the areas of nanotechnologies, biotechnologies, aerodynamics, hydrodynamics, stress tests, information processing (data mining,artificial learning, semantic analysis, etc.), complex systems and systemic analysis, man/machineinterface and interaction, social networks, neural networks, and so on. As one can see, the list is longand the applications numerous.Thus, we are currently witnessing the modeling and simulations of behavioral strategies of agents or players interacting with their environment. This kind of problem, which can be considered a part of thegame theory with N players, when N is very big, has very concrete applications in the management of commodities, social networks, conflicts, the modeling of crowd movements, and economics.
From Mathematics to Innovative Use
The experience of Thierry Lorho, the designer of Artificial Intelligence system Globe Expert withinInternational Focus, perfectly illustrates this perspective. Globe Expert was selected by DARPA(Defense Advanced Research Projects Agency) in the framework of its prospective on the futuregeopolitical background of human expansion into outer space. This AI system also serves as a basis of analysis and simulation for the research program
“Dynamics of Emotions and Fear in Conflicts and Reconstruction,”
jointly carried out by the Graduate Institute of International and DevelopmentStudies (Geneva) and Lausanne Technology Institute (Behavioral Genetics Laboratory). In the beginning, Globe Expert was designed for applications focused on the major changes of the 21
century: geopolitical and economic, environmental, scientific, technological, and societal. Here,modeling complexity and anticipating uncertainty are at stake by comprehending complex and sudden phenomena, which occur on the edge of reality and are also called “black swans” or “dragon kings.”Understanding Artificial Intelligence can seem difficult. Indeed, it involves multiple disciplines and, beyond expert eyes on particular topics, only the broad vision of a generalist, addressing the wholefields and establishing relations between them, is able to go to task.Artificial Intelligence is primarily and basically concerned with signals more than language andsemantics. In order to design Globe Expert, the following processes and their combinations wereneeded:1.
Understanding information theory and the fundamental definition of entropy as given byClaude Shannon;2.
Building on thermodynamics and above all the second law, important to Boltzmann, whichdefines the entropy notion of a complex system;3.
Broadening this comprehension to the quantum theory of information and subtleties of statessuperposition, coherences and decoherences;4.
Further broadening this horizon to singular phenomena where information, in its mostfundamental form, is at the heart of modern physics (disconcerting concept of multiverse,holographic universe, quantum theory of black holes and entropy limits of Bekenstein-Hawking);5.
Focusing attention on quantum theories of decision and games, graph theory, complexitymeasure, distances calculation;6.
Mastering different models; environmental, neural, econometric and others (Lotka-Volterraequations, PSO, perceptrons, Hash model…);7.
Mastering the semantic issues, structures of language (Zipf-Mandelbrot law, multilingualismissues, global or structured approach of comprehension);8.
Addressing the human paradox in its infinite complexity in order to model it through aquantum approach of decision (for instance Allais or Ellsberg paradox).Globe Expert’s consistency is driven by its practical applications and its ability to replace thecomplicated visible by simple invisible. First of all, the fundamentals of information theory are practically applied. Then, a clusters’ system integrates information in analytical frameworks. This