Research map / Video / Video generation
GenFlowRL: Shaping Rewards with Generative Object-Centric Flow in Visual Reinforcement Learning
Kelin Yu, Sheng Zhang, Harshit Soora, Furong Huang, Heng Huang, Pratap Tokekar, Ruohan Gao
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use GenFlowRL instead of Im2Flow2Act imitation policy | GenFlowRL · Im2Flow2Act Pivoting (demo-conditioned) | success rate (%): 60 -> 90 | Improved |
| use GenFlowRL instead of Im2Flow2Act imitation policy | GenFlowRL · Im2Flow2Act Folding (language-conditioned) | success rate (%): 35 -> 80 | Improved |
| use GenFlowRL instead of Im2Flow2Act imitation policy | GenFlowRL · Im2Flow2Act Pivoting (language-conditioned) | success rate (%): 45 -> 85 | Improved |
| add large Gaussian noise (x4) to generated flow (ablation) | GenFlowRL · Im2Flow2Act Folding (language-conditioned) | success rate (%): 80 -> 75 | No clear effect |
| add large drift noise (Gauss x2 Drift x2) to generated flow (ablation) | GenFlowRL · Im2Flow2Act Pouring (language-conditioned) | success rate (%): 95 -> 75 | Hurt |
| add large drift noise (Gauss x2 Drift x2) to generated flow (ablation) | GenFlowRL · Im2Flow2Act Folding (language-conditioned) | success rate (%): 80 -> 65 | Hurt |
| encode keypoints with delta-flow instead of MLP (ablation) | GenFlowRL · PickNPlace, Pouring, Pivoting (RL training) | success rate | Improved |
| add initial 3D centroids to policy input (ablation) | GenFlowRL · PickNPlace, Pouring, Pivoting (RL training) | success rate | Improved |
| use 128 keypoints instead of 64 or 32 (ablation) | GenFlowRL · PickNPlace, Pouring, Pivoting (RL training) | success rate | No clear effect |
| use delta-flow reward instead of VIPER, Diffusion Reward, RND or sparse reward | GenFlowRL · MetaWorld (5 hard tasks) | training success rate | Improved |
Closest papers by meaning
- FLIP: Flow-Centric Generative Planning as General-Purpose Manipulation World Model (ICLR 2025 · similarity 0.89)
- VideoVLA: Video Generators Can Be Generalizable Robot Manipulators (NeurIPS 2025 · similarity 0.87)
- Generative Control as Optimization: Time Unconditional Flow Matching for Adaptive and Robust Robotic Control (NeurIPS 2026 · similarity 0.87)
- GenPO: Generative Diffusion Models Meet On-Policy Reinforcement Learning (NeurIPS 2025 · similarity 0.87)
- GraphMimic: Graph-to-Graphs Generative Modeling from Videos for Policy Learning (CVPR 2025 · similarity 0.87)
- ReinFlow: Fine-tuning Flow Matching Policy with Online Reinforcement Learning (NeurIPS 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
