We propose and analyze an alternate approach to off-policy multi-step temporal difference learning, in which off-policy returns are corrected with the current Q-function in terms of rewards, rather than with the target policy in terms of transition probabilities. We prove that such approximate corrections are sufficient for off-policy convergence both in policy evaluation and control, provided certain conditions. These conditions relate the distance between the target and behavior policies, the eligibility trace parameter and the discount factor, and formalize an underlying tradeoff in off-policy TD(λ). We illustrate this theoretical relationship empirically on a continuous-state control task.
CITATION STYLE
Harutyunyan, A., Bellemare, M. G., Stepleton, T., & Munos, R. (2016). Q(λ) with off-policy corrections. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9925 LNAI, pp. 305–320). Springer Verlag. https://doi.org/10.1007/978-3-319-46379-7_21
Mendeley helps you to discover research relevant for your work.