About me
I am an MSCA postdoctoral fellow at the KU Leuven (Belgium), where I perform research at the intersection of deep learning, computational neuroscience and cognitive science. Until February 2022, I was a postdoc at the Baylor College of Medicine (Houston, Texas) and a visiting researcher at the University of Cambridge.
The main goal of my research is to understand the computational principles underlying the cognitive skill of continual learning. I do this through a combination of conceptual analysis, computational modelling, deep neural network implementations and collaborations with experimental labs. My contributions include proposing the influential “three scenarios” framework for continual learning, providing a proof-of-principle demonstration that generative classification is a promising strategy for class-incremental learning, and identifying the stability gap—an intriguing phenomenon in which deep neural networks suffer substantial but temporary forgetting when starting to learn something new.
Another research direction I am interested in is using insights and intuitions from neuroscience to make the behavior of deep neural networks more human-like. For example, I developed the brain-inspired replay method, which alleviates catastrophic forgetting in deep neural networks by replaying self-generated, abstract memory representations.
Previously, for my award-winning PhD in Neuroscience (University of Oxford), I used optogenetics and electrophysiological recordings in mice to study the role of replay in memory consolidation in the brain. Before that I obtained degrees in Statistics (Master, UC Berkeley) and Econometrics (Bachelor, Erasmus University Rotterdam).