University of Bristol

Carl Henrik Ek

I am interested in learning factorised models where the latent representation encodes some form of structure. In specific I have focused on learning models in the scenario when data comes from several different view. I have been working on a large range of different applications, from computer vision and robotics to computational biology. Recently I have become interested in representations of human and animal behavior, trying to create models which allows us to understand how our mental state effects our behavior and trying to discover structures among individuals interacting. I try to focus on applications where data is scares such that understanding the data is necessary which means that we require “real” models. The main methodology is Bayesian non-parametrics in specific Gaussian Processes.

Short Bio
Before joining Bristol I was an Assistant Professor in Machine Learning at the Royal Institute of Technology (KTH) in Stockholm. I was a postdoctoral research at University of California at Berkeley and I did my PhD research at Oxford Brookes University. During my PhD I was also a member of the machine learning group at University of Sheffield and University of Manchester. I hold a MEng degree in Vehicle Engineering and a docent degree in Machine Learning from the Royal Institute of Technology in Stockholm.