Genie Centurion: Accelerating Scalable Real-World Robot Training with Human Rewind-and-Refine Guidance
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915737258426368 |
|---|---|
| author | Wang, Wenhao Song, Jianheng Liu, Chiming Ma, Jiayao Feng, Siyuan Wang, Jingyuan Jiang, Yuxin Chen, Kylin Zhan, Sikang Wang, Yi Meng, Tong Shi, Modi He, Xindong Ren, Guanghui Yang, Yang Yao, Maoqing |
| author_facet | Wang, Wenhao Song, Jianheng Liu, Chiming Ma, Jiayao Feng, Siyuan Wang, Jingyuan Jiang, Yuxin Chen, Kylin Zhan, Sikang Wang, Yi Meng, Tong Shi, Modi He, Xindong Ren, Guanghui Yang, Yang Yao, Maoqing |
| contents | While Vision-Language-Action (VLA) models show strong generalizability in various tasks, real-world deployment of robotic policy still requires large-scale, high-quality human expert demonstrations. However, data collection via human teleoperation requires continuous operator attention, which is costly, hard to scale. To address this, we propose Genie Centurion (GCENT), a scalable and general data collection paradigm based on human rewind-and-refine guidance, enabling robots' interactive learning in deployment. GCENT starts at an imperfect policy and improves over time. When the robot execution failures occur, GCENT allows robots to revert to a previous state with a rewind mechanism, after which a teleoperator provides corrective demonstrations to refine the policy. This framework supports a one-human-to-many-robots supervision scheme with a Task Sentinel module, which autonomously predicts task success and solicits human intervention when necessary. Empirical results show that GCENT achieves up to 40% higher task success rates than state-of-the-art data collection methods, and reaches comparable performance using less than half the data in long-horizon and precise tasks. We also quantify the data yield-to-effort ratio under multi-robot scenarios, demonstrating GCENT's potential for scalable and cost-efficient robot policy training in real-world environments. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_18793 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Genie Centurion: Accelerating Scalable Real-World Robot Training with Human Rewind-and-Refine Guidance Wang, Wenhao Song, Jianheng Liu, Chiming Ma, Jiayao Feng, Siyuan Wang, Jingyuan Jiang, Yuxin Chen, Kylin Zhan, Sikang Wang, Yi Meng, Tong Shi, Modi He, Xindong Ren, Guanghui Yang, Yang Yao, Maoqing Robotics While Vision-Language-Action (VLA) models show strong generalizability in various tasks, real-world deployment of robotic policy still requires large-scale, high-quality human expert demonstrations. However, data collection via human teleoperation requires continuous operator attention, which is costly, hard to scale. To address this, we propose Genie Centurion (GCENT), a scalable and general data collection paradigm based on human rewind-and-refine guidance, enabling robots' interactive learning in deployment. GCENT starts at an imperfect policy and improves over time. When the robot execution failures occur, GCENT allows robots to revert to a previous state with a rewind mechanism, after which a teleoperator provides corrective demonstrations to refine the policy. This framework supports a one-human-to-many-robots supervision scheme with a Task Sentinel module, which autonomously predicts task success and solicits human intervention when necessary. Empirical results show that GCENT achieves up to 40% higher task success rates than state-of-the-art data collection methods, and reaches comparable performance using less than half the data in long-horizon and precise tasks. We also quantify the data yield-to-effort ratio under multi-robot scenarios, demonstrating GCENT's potential for scalable and cost-efficient robot policy training in real-world environments. |
| title | Genie Centurion: Accelerating Scalable Real-World Robot Training with Human Rewind-and-Refine Guidance |
| topic | Robotics |
| url | https://arxiv.org/abs/2505.18793 |