Kyriacos Shiarlis

Kyriacos Shiarlis

DPhil Student

My PhD is focused on Learning from Demonstration (LfD), applied primarily to robotics. So far, I have developed methods that allow agents to learn from failed demonstrations, learn cost functions for sample based planners and acquire social behaviours using deep learning. I believe that LfD will play a key role in Reinforcement Learning and AI, especially in settings where cost function functions are either too sparse or too difficult to define.

Before starting a PhD I graduated with an BEng in Mechanical Engineering and an MEng in Information Engineering from the University of Cambridge.