¹⁸(Sutton, 1991; Sutton and Barto, 2018). See also Mattar and Daw (2018) for a theoretical unification of replay and planning, and Kurth-Nelson et al. (2023) for neuroscience results and theory on how replay is more than just repeating past episodes. For a popular-science account focusing on the utility of regret, see Pink (2022).