Kiet Bennema ten Brinke

Creating noise from data is trivial; creating data from noise is generative modelling 

As PhD Candidate in the Machine Learning for Physical Sciences group I work on deep generative modelling solutions for the simulation of stochastical dynamical systems. My focus lies on designing and implementing efficient deep learning architectures for spatiotemporal data of several data domains. By using new generative methods and looking at symmetries present in the data our group create models that achieve higher simulation fidelity while wasting less computational resources.