FlowRL: A Taxonomy and Modular Framework for Reinforcement Learning with Diffusion Policies Paper • 2603.27450 • Published Mar 29