Research map / Vision-language and multimodal / Robot action and navigation
Boosting Vision-Language-Action Finetuning with Feasible Action Neighborhood Prior
Haochen Niu, Kanyu Zhang, Shuyu Yin, Qinghai Guo, Peilin Liu, Fei Wen
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 FAN-guided Gaussian regularizer to SFT | OpenVLA · ManiSkill in-distribution | success rate (%): 78.1 -> 89.8 | Improved |
| add FAN-guided regularizer to SFT | OpenVLA · ManiSkill OOD average | success rate (%): 58.1 -> 63.3 | Improved |
| add FAN-guided regularizer to PPO | OpenVLA · ManiSkill OOD average | success rate (%): 81.9 -> 88.1 | Improved |
| add FAN-guided regularizer to PPO | OpenVLA · ManiSkill in-distribution | success rate (%): 95.9 -> 97.4 | Improved |
| add FAN-guided regularizer to PPO | OpenVLA-OFT · ManiSkill OOD average | success rate (%): 63.3 -> 71.2 | Improved |
| add FAN-guided regularizer to PPO | OpenVLA-OFT · ManiSkill in-distribution | success rate (%): 92.3 -> 97.3 | Improved |
| add FAN-guided regularizer to PPO | OpenVLA · ManiSkill training sample efficiency | training steps to 90% success: about one-third of baseline steps | Improved |
| add FAN-guided regularizer to SFT | OpenVLA · LIBERO-Spatial | success rate (%): 84.7 -> 87.2 | Improved |
| add FAN-guided regularizer to SFT | OpenVLA-OFT · LIBERO-Spatial | success rate (%): 95.2 -> 98.8 | Improved |
| add FAN-guided regularizer to SFT | OpenVLA · LIBERO-Spatial with position perturbation (0.05, 0.05) | success rate: 0.24 -> 0.36 | Improved |
| add FAN-guided regularizer to SFT | OpenVLA · Real robot perturbed manipulator pose (Task-3) | successes / 30: 7/30 -> 17/30 | Improved |
| add FAN-guided regularizer to SFT | OpenVLA · Real robot standard task (Task-1) | successes / 30: 19/30 -> 22/30 | Improved |
Closest papers by meaning
- ActionUNet: Improving Robustness of VLA Models with Efficient Multi-scale Fine-tuning (NeurIPS 2026 · similarity 0.91)
- PriorVLA: Prior-Preserving Adaptation for Vision-Language-Action Models (NeurIPS 2026 · similarity 0.91)
- Latent Action Pretraining from Videos (ICLR 2025 · similarity 0.90)
- ElegantVLA: Learning When to Think for Efficient Vision-Language-Action Models (NeurIPS 2026 · similarity 0.90)
- Beyond World-Frame Action Heads: Motion-Centric Action Frames for Vision-Language-Action Models (NeurIPS 2026 · similarity 0.90)
- NS-VLA: Towards Neuro-Symbolic Vision-Language-Action Models (NeurIPS 2026 · similarity 0.90)
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.
