Peter Dekker
Pleinlaan 9, 1050 Brussels, 3rd floor

Peter Dekker

PhD Student


I am a PhD student at the AI lab of the Vrije Universiteit Brussel, with prof. Bart de Boer. I use agent-based models to investigate how individual behaviour can amount to collective behaviour, and how individuals can be stimulated to make choices that lead to desirable collective behaviour. As a case study of collective behaviour, I study language change, a domain for which an abundance of data is available, on the individual and collective level. In a language, speakers would like to minimize effort (self-interest), while still getting their message across (collectively desirable behaviour). I develop agent-based models of language change relying on general principles, which can be applied to other collective systems, such as organizational, societal or traffic systems. My research is funded by the Flanders AI Programme, challenge Multi-agent Collaborative AI.

My general research interest is the application of computational methods to answer questions about language change and language ancestry. I studied Artificial Intelligence at Utrecht University (BSc) and the University of Amsterdam (MSc), where I wrote my MSc thesis Reconstructing language ancestry by performing word prediction with neural networks. I previously worked as a systems developer at the Dutch Language Institute, on collection and processing of language data, among others via crowdsourcing.