Nicholas Roy

I work as a graduate data scientist in economic consulting (views my own). Prior to this, I completed the MSc in Statistical Science at Oxford.

My aim with this blog is to record and explain various concepts that I encounter in machine learning research, with more emphasis on intuition than rigour. I hope that it will be useful to others exploring related ideas.

My research interests lie primarily in model-based reinforcement learning and sim-to-real transfer. I am especially interested in how an agent should act in order to improve its model of the world. Previously, I approached this via model misspecification in Bayesian experimental design.