RallyNet, a hierarchical offline imitation learning model, adeptly imitates badminton player behaviors by leveraging experiential contexts and geometric Brownian motion.
