FWO WOG - Guiding networked societies, linking data science and modelling

Networks of interconnected autonomous computing entities more and more support our society, interacting and influencing each other in complex and unforeseen ways. Examples are smart grids, intelligent traffic lights, logistics and voluntary peer­to­peer clouds as well as socio­technical systems or more generally the Internet of Things. Understanding the characteristics and dynamics of these systems both at the local and global scale is crucial in order to be able to guide such systems to desirable states.

The partners participating in this WOG proposal each study crucial features of such complex systems, or they are experts in related fields that offer complementary techniques to analyze the massive data that is generated by them. Bringing these orthogonal fields of expertise together in a scientific research community, promises to give great opportunity for cross-fertilization and the development of novel analysis and control techniques.

Today, the actors in these networks are equipped with sensors that constantly gather information or generate measurements that can easily be collected. These measurements can be enriched with meta-information, allowing for intelligent techniques to retrieve the patterns that drive the actors as well as the overall phenomena captured by the technology. This data is not only large (big data) but it also brings new challenges as it is also heterogeneous, noisy and is generated at a high speed in an asynchronous manner.

Moreover, these patterns can be used to design models of the system that can then be investigated in terms of the system’s (as well as the individual actor’s) resistance against perturbations or malleability towards well-defined advantageous states, where advantage is defined by societal needs or technological efficiency. As these networks are very hard to model, therefore Model Free Learning techniques, (MFL) such as reinforcement learning, are very promising in this context. Since the scale of these networks is very large, there are interesting open questions on how to organize a MFL approach, especially on how the MFL can benefit from the data analysis in order to improve the learning in terms of speed and quality of the obtained solution.

As mentioned above, the systems we will study are heterogeneous, as they often combine technology with human preferences and behaviors. Obviously humans cannot be controlled in the same way as devices. Nonetheless, guidance to avoid particular outcomes is possible, given the right incentives. Like the MFL mentioned earlier, population-based modeling approaches like evolutionary dynamics can be exploited here together with behavioral experiments to determine the behavior changing mechanisms relevant for the problem in question, which can be tested in academic and real-world scenarios. Pattern mining techniques can in turn be used to construct the actors in such systems as they identify the emergence of the behavioral change in the real-world applications.

In conclusion, this WOG aims to bring together a unique set of expertise to study these new systems that are going to play a crucial role in our digital society in the coming decades.

The involved Flemish community partners are the following:

AI-lab, Vrije Universiteit Brusel
Data Science Lab, Universiteit Gent
Onderzoeksgroep DTAI, KU Leuven
Advanced Database Research and Modelling, Universiteit Antwerpen
iMinds-DistriNet, KU Leuven

The involved partners not part of the Flemish community are:

Intelligent Systems Laboratory, University of Bristol
Intelligent Agents and Synthetic Characters Group (GAIPS) & ATP-group, INESC-ID, Instituto de Engenharia de Sistemas e Computadores - Investigação e Desenvolvimento (INESC-ID), Instituto Superior Técnico, University of Lisbon
Artificial Intelligence section, University of Liverpool
Algorithmic Data Analytics Lab, The ISI Foundation