About me

I’m a PhD student at the Amsterdam Machine Learning Lab supervised by Jan-Willem van de Meent. As student in the UvA-Bosch Delta Lab I am also supervised by Sebastian Gerwinn. I am interested in generative models, with a focus on flow-based methods, distillation methods, and graph generation. Previously, I wrote my Master’s thesis on Autoencoders within Flows (AEFs) under the supervision of Luca Ambrogioni.

I obtained a Master’s degree in Computing Science from the Radboud University Nijmegen with a specialization in Data Science.

Publications/preprints

• Daan Roos^*, Oscar Davis^*, Floor Eijkelboom^*, Michael M. Bronstein, Max Welling, Ismail Ilkan Ceylan, Luca Ambrogioni, Jan-Willem van de Meent. Categorical Flow Maps. Preprint [Paper]
• Daan Roos, Sebastian Gerwinn, Jan-Willem van de Meent, Sara Magliacane. Improving Local Explainability By Learning Causal Graphs From Data. TMLR. [Paper]
• Gianluigi Silvestri^*, Daan Roos^*, Luca Ambrogioni. Deterministic training of generative autoencoders using invertible layers. International Conference on Learning Representations 2023. [Paper]