Junior Lecturer / PhD Candidate for Bayesian Dynamic Control at the Donders Centre for Cognition

Are you an aspiring researcher and would you like to start your academic career off well-prepared? As a PhD candidate on the topic of Bayesian dynamic control, you will not only develop state-of-the-art machine learning algorithms, but also teach these skills to new students in the AI programme. This way, you will be able to obtain both your PhD and teaching qualifications in one position.
Dynamical systems are prevalent in many academic and engineering domains, such as healthcare, climate science, economics, epidemiology and neuroscience. The dynamics of these systems are often described by a set of governing equations, and academia has produced several advanced algorithms to identify these from observational data. In this project, we close the loop between systems identification on the one hand and optimal control on the other, in order to optimally manipulate a dynamical system, and simultaneously adjust our beliefs about the dynamical system based on how it responds to our actions. We aim to develop AI agents capable of extracting information from their environment in order to solve complex tasks. This will involve the integration of approximate Bayesian inference techniques with state-of-the-art machine learning methods.
In this unique position, teaching plays a crucial role. Not only will you be developing state-of-the-art machine learning algorithms (60% of your time), you will also be teaching the skills to do so to new students in the AI programme (40% of your time). This allows you to obtain both your PhD and teaching qualifications in one position, making you well-prepared for an academic career.