Abstract page for arXiv paper 2611.14599: StructRL: Model-Free Reinforcement Learning with Replay Graphs