AMLab: The Amsterdam Machine Learning Lab (AMLab) conducts research in the area of large scale modeling of complex data sources. This includes the development of new methods for probabilistic graphical models and nonparametric Bayesian models, the development of faster (approximate) inference and learning methods, deep learning, causal inference, reinforcement learning and multi-agent systems and the application of all of the above to large scale data domains in science and industry (“Big Data problems”).
Within AMLAB my group is mainly interested in deep learning, graphical models and large scale inference and learning. The following projects are currently actively studied:
-Variational Auto-Encoders (with Durk Kingma).
-Transformation equivariant convolutional neural networks (with Taco Cohen)
-Likelihood free simulator models or Approximate Bayesian Computation (with Ted Meeds)
-Deep neural networks for genomics (with Ted Meeds)
-Deep learning for radio astronomy (with Patrick Putzky)
-Deep unsupervised and sequence modeling (with Karen Ullrich)
-Bayesian deep learning (with Christos Louizos)
-Deep learning for medical imaging (with Tameem Adel)
-Network analysis (with Thomas Kipf)
-Knowledge graphs and deep learning (with Wenzhe Li)
QUVA Lab: QUVA conducts research in deep learning for computer vision.
Within QUVA I supervise four students working on:
-Spiking neural networks (with Peter O’Connor)
-Privacy preserving deep learning (with Mijung Park)
-Bayesian distributed deep learning (with Matthias Reisser)
-Bayesian optimization of deep networks (with ChangYong Oh)