L.R. Engwegen

My research interests are in deep reinforcement learning and (computational) neuroscience. As part of the BIOLab, I am doing research in AI with possible applications in neuroscience. More specifically, I am investigating deep reinforcement learning methods that can analyze and control activity in networks of biological neurons. A key aspect of my research is the generalizability of such methods in a partially observable domain. We are developing recurrent and transformer-based architectures to overcome the challenges in such setting.

On the application side, I am interested in modelling biological neuronal systems to (1) better understand their dynamics, and (2) apply deep RL methods to control their activity. The aim is to close to gap with clinical and research applications in which we want to, for example, classify diseased neurons or restore healthy behavior. Additionally, I am interested in classical generalization and partial observability problems in RL.