Daswani, Mayank; Sunehag, Peter; Hutter, Marcus
We extend the Q-learning algorithm from the Markov Decision Process setting to problems where observations are non-Markov and do not reveal the full state of the world i.e. to POMDPs. We do this in a natural manner by adding l0 regularisation to the pathwise squared Q-learning objective function and then optimise this over both a choice of map from history to states and the resulting MDP parameters. The optimisation procedure involves a stochastic search over the map class nested with classical...[Show more]
Items in Open Research are protected by copyright, with all rights reserved, unless otherwise indicated.