Z. Wang, “One-Step Generative Policies with Q-Learning: A Reformulation of MeanFlow”, AAAI, vol. 40, no. 31, pp. 26751–26759, Mar. 2026.