A curious robot: An explorative-exploitive inference algorithm
Kim Steenstrup Pedersen, Peter Johansen · Research at the University of Copenhagen (University of Copenhagen) · 2007
We propose a sequential learning algorithm with a focus on robot control. It is initialised by a teacher who directs the robot through a series of example solutions of a problem. Left alone, the control chooses its next action by prediction based on a variable order Markov chain model selected to minimise a MDL criterion based on generalised code length La of the past robot-environment interaction. The user specifies the parameter a and as a result, the robot can be directed towards exploratory behaviour if confidence in the teacher is low (a0). The novelty of the proposed method lies in the use of generalised code length in the MDL model selection criterion.