I'm a Ph.D. student at Uppsala University primarily interested in reinforcement learning. Before that, I was an AI Research Engineer at Peltarion AI Research where I went after graduating from Chalmers University of Technology.
An extended abstract, Bayesian Linear Regression on Deep Representations, based on my thesis, was accepted to the 4th workshop on Bayesian Deep Learning at NeurIPS 2019 in Vancouver, Canada, and I'm excited to be able to attend the conference!
I just presented and published my Master thesis on Uncertainty-Aware Models for Deep Reinforcement Learning. I use Bayesian linear regression on the representation from the last layer of a neural network to get epistemic and heteroscedastic aleatoric uncertainty estimates.