Public defence in Computer Science, M.Sc. (Tech.) Tuomas Takko

Title of the thesis: Data-driven modelling of human behaviour with complex networks

Doctoral student: Tuomas Takko
Opponent: Professor Stefan Thurner, Medical University of Vienna, Austria
Custos: Professor Jouko Lampinen, Aalto University School of Science, Department of Computer Science

Data-driven modelling of human behaviour with complex networks 

There is more information and data available today than ever before. Text-based reports and news, mobile phone applications and location information, as well as controlled experiments on virtual platforms enable conducting new data-driven research on the mechanisms and phenomena related to human activities. The complex and multidimensional human behaviour can be approached in a computational manner using modelling. In order to gain understanding of the phenomenon being investigated, the modelling methodology needs to be sufficiently accurate in representing the interactions and sufficiently simple to be interpretable. In this doctoral dissertation data-driven modelling of human behaviour was conducted in three timely contexts: decision making in game experiments, mobility and exposure during the Covid-19 pandemic and in cyber incidents targeting organizations. The objective of the research is to understand the phenomena through data-driven computational models and to depict the full process of modelling for the data and the investigated system. In the research of decision making of humans cooperating with autonomous software agents, the human perception of risk was found to be near optimal and the group consistency of humans and agents was found to have an effect on the strategies and outcomes. Results from the research provide new information on human decision making in cooperative games and hybrid systems. The research on human mobility and exposure during the pandemic showed a significant change from the pre-pandemic's nearly stable regime as well as the effect of the government-imposed restrictions. The results can be utilized in evaluating and modelling the effect of the restrictions on exposure from non-invasive aggregated data. The final study of the dissertation presented a framework for mining open textual reports and constructing a knowledge graph for analysis and computational methods. The resulting framework can be beneficial to experts in the field by constructing a broad high-level model of the incidents and the similarities of the targeted organizations.

Key words: data-driven modelling, human behaviour, complex systems, complex networks

Thesis available for public display 10 days prior to the defence at: https://aaltodoc.aalto.fi/doc_public/eonly/riiputus 

Doctoral theses of the School of Science: https://aaltodoc.aalto.fi/handle/123456789/52