We aim at reconstructing the connectivity networks from data and to predict the sudden changes of their behavior. Thereby we will be able to develop strategies to avoid malfunctioning.
The first step is develop a mathematical theory to describe emergent behavior in complex networks of nonlinear dynamical systems. These complex systems have a wild globally behavior making relevant phenomena such as collective dynamics to take place over finite time scales. As a consequence, these systems cannot be tackled by standard tools. The ReconNeCT Project fills this gap by developing a theory for emergent phenomena in complex networks. This theory opens enormous possibilities such as predicting critical transitions in complex systems and reconstructing the network dynamics from data.
1) For each connectivity level, this theory will provide a low-dimensional description over ﬁnite time scales. Thereby, it allows for a description of collective phenomena in terms of the network structure.
2) The above objectives provide eﬀect models over connectivity layers. Hence, it oﬀers a way to reverse engineer the problem and obtain a model from data. Consequently, opening the possibility to predict critical transitions from data which would be impossible otherwise.
We will apply this theory to neuron networks and smart cities sensor data to predict the critical transitions and to help preventing catastrophes.