Peter Dekker

Pleinlaan 9, 1050 Brussels, 3rd floor

Peter Dekker

PhD Student


I am a PhD student in agent-based models of language change, in Bart de Boer‘s ABACUS group at the AI lab of the Vrije Universiteit Brussel. In my PhD project (start: 2019), I use agent-based models to study the social, cognitive and language-specific factors that drive language change. I apply my models to case studies from the languages of the world. I am currently studying the relationship between morphological simplification and Papuan-Austronesian language contact in Eastern Indonesia. In addition to agent-based models, I draw upon techniques from deep learning and bioinformatics. Moreover, together with colleagues, I am exploring the use of agent-based models for language preservation. My research is supported by an FWO PhD Fellowship Identifying drivers of language change using neural agent-based models. I was previously funded by the Flanders AI Program.

My general interest is the use of computational methods to study language change, from a bottom-up (agent-based modelling, neural networks, crowdsourcing) and a top-down (phylogenetic reconstruction, computational historical linguistics) perspective. I studied Artificial Intelligence at the University of Utrecht and the University of Amsterdam. I previously worked as a systems developer at the Dutch Language Institute.