#ReinforcementLearning in Presence of Discrete Markovian Context Evolution - @UCL Discovery