Dr Calum Imrie
With huge strides being made with robotic applications and their capabilities, it is widely considered that it will ultimately be commonplace for robots to fully take over certain responsibilities. It is conceivable, for example, that an entire factory could be governed and operated by robotics. For scenarios such as this to be realised, however, the robots in question must be fully autonomous, and a possible way of realising this is to replicate human learning behaviour. Confidence, a firm trust that a chosen course of action is the best or most effective, allows a person to identify if they are capable of an action, but more than that, it manipulates the person's attitude towards learning tasks. This project's aim is to identify if a higher level concept of confidence can be used as a self-learning tool to help gain a level of autonomy for robotic applications.
Graduated with a BSc in Artificial Intelligence and Computer Science from the University of Edinburgh. Current interests are different learning behaviours applied to robotic applications, which includes both single agents and multi-agents interactions (for example swarm robotics) given the same learning behaviour.