Transfer Q-learning

Elynn Chen, Sai Li, Michael I. Jordan

Time-inhomogeneous finite-horizon Markov decision processes (MDP) are frequently employed to model decision-making in dynamic treatment regimes and other statistical reinforcement…