Research map / Reinforcement learning / Imitation learning
CE-Nav: Flow-Guided Reinforcement Refinement for Cross-Embodiment Local Navigation
Kai Yang, Tianlin Zhang, Zhengbo Wang, Zedong Chu, Xiaolong Wu, Yang Cai, Mu Xu
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 |
|---|---|---|---|
| add expert guidance (VelFlow prior) vs RL from scratch (ablation) | CE-Nav · Obstacle forest N_o=500, Go2 | success rate: 0.56 -> 0.83 | Improved |
| add expert guidance, training time (ablation) | CE-Nav · Go2 adaptation | ETT (h): 52 -> 6 | Improved |
| use VelFlow guidance instead of MLP regression guidance (ablation) | CE-Nav · Obstacle forest N_o=500, Go2 | success rate: 0.28 -> 0.83 | Improved |
| regression guidance vs no guidance (ablation) | CE-Nav regr-rl · Obstacle forest N_o=500, Go2 | success rate: 0.56 -> 0.28 | Hurt |
| use VelFlow instead of diffusion policy as guide (ablation) | CE-Nav · Obstacle forest N_o=500, Go2 | success rate: 0.77 -> 0.83 | Improved |
| add RL refinement to the VelFlow expert (ablation) | CE-Nav · Obstacle forest N_o=300, Go2 | success rate: 0.01 -> 0.84 | Improved |
| anneal guidance weight instead of constant 0.5 (ablation) | CE-Nav · Obstacle forest N_o=500, Go2 | success rate: 0.73 -> 0.83 | Improved |
| use CE-Nav instead of NavRL | CE-Nav · Obstacle forest avg, Go2 | mean success rate: 0.6925 -> 0.8575 | Improved |
| transfer to other embodiments with same expert | CE-Nav · Five robots | mean success rate: 0.7450 to 0.8600 across robots | No clear effect |
Closest papers by meaning
- ReinFlow: Fine-tuning Flow Matching Policy with Online Reinforcement Learning (NeurIPS 2025 · similarity 0.88)
- Implicit Drifting Policy: One-Step Action Generation via Conditional Expert Geometry (NeurIPS 2026 · similarity 0.87)
- Reasoning in Visual Navigation of End-to-end Trained Agents: A Dynamical Systems Approach (CVPR 2025 · similarity 0.87)
- C-NAV: Towards Self-Evolving Continual Object Navigation in Open World (NeurIPS 2025 · similarity 0.87)
- Behavioral Exploration: Learning to Explore via In-Context Adaptation (ICML 2025 · similarity 0.87)
- PEAC: Unsupervised Pre-training for Cross-Embodiment Reinforcement Learning (NeurIPS 2024 · similarity 0.86)
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.
