Intentionality looks to be a main distinguishing feature of the motion in any metric space that distinguishes animate from inanimate beings. Escaping the vagueness of this sentence, we characterize this feature at three levels through the modeling of specific phenomena:
1) from a kinematic perspective we analyze the flocking behavior in nature, ranging from the albatross flights to the mobility of people sharing a common goal, till epidemic phenomena. All the related trajectories denote a symmetry breaking in respect to Brownian motion that is formally captured by us through a new variant of the Pareto distribution law.
2) from a dynamic perspective we introduce a new neural network paradigm where neurons are allowed to move to find their best reciprocal positions with the aim of efficiently transmitting information. This gives rise to dynamics – in an extended metric space where neuron Euclidean coordinates merge with connection weights – that are ruled by a generalized Hamiltonian made of kinetic and cognitive components.
3) moving to a large scale we instantiate the above network in terms of social networks, where we analyze the clustering effects of these dynamics in view of forming interest groups with suitable properties. We study these effects in a rather abstract way by considering the two template instances of the “learning by gossips paradigm” and a cooperative training to recognize the digits of the MNIST database.
[Biography] Bruno Apolloni is full professor in Computer Science at the Department of Computer Science of the Milan University, Italy.?His main research interests are in the frontier area between probability, mathematical statistics and computer science, with special regard to statistical bases of learning, subsymbolic and symbolic learning processes, granular computing, analysis of biomedical data, and modeling of dynamical processes in biology.He introduced the Algorithmic Inference approach in statistics as a conceptual and methodological tool to solve modern computational learning problems with the massive use of computers. In particular, it provides a unifying theoretical framework to the various data analysis and management disciplines converging under the granular computing heading. He also introduced some non-markovian processes to model intentionality in a wide range of biological systems ranging from bacteria colonies to social communities.
Apolloni is head of the Neural Networks Research Laboratory (LAREN http://laren.dsi.unimi.it) at the University of Milan, past President of the Italian Society for Neural Networks (SIREN, http://siren.dsi.unimi.it), and member of the European NeuralNetwork Society (ENNS,http://www.e-nns.org/) board. He is a member of the editorial board of many journals in the field, among which: Neural Networks, Neurocomputing, International Journal of Hybrid Intelligent Systems, International Journal of Information and Communication Technology, and International Journal of Computational Intelligence Studies.He is Scientific Leader in national and international research projects. He published around 160 papers.