A Self-Correcting Vision-Language-Action Model for Fast and Slow System Manipulation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917960412561408 |
|---|---|
| author | Li, Chenxuan Liu, Jiaming Wang, Guanqun Li, Xiaoqi Chen, Sixiang Heng, Liang Xiong, Chuyan Ge, Jiaxin Zhang, Renrui Zhou, Kaichen Zhang, Shanghang |
| author_facet | Li, Chenxuan Liu, Jiaming Wang, Guanqun Li, Xiaoqi Chen, Sixiang Heng, Liang Xiong, Chuyan Ge, Jiaxin Zhang, Renrui Zhou, Kaichen Zhang, Shanghang |
| contents | Recently, some studies have integrated Multimodal Large Language Models into robotic manipulation, constructing vision-language-action models (VLAs) to interpret multimodal information and predict SE(3) poses. While VLAs have shown promising progress, they may suffer from failures when faced with novel and complex tasks. To emulate human-like reasoning for more robust manipulation, we propose the self-corrected (SC-)VLA framework, which integrates fast system for directly predicting actions and slow system for reflecting on failed actions within a single VLA policy. For the fast system, we incorporate parameter-efficient fine-tuning to equip the model with pose prediction capabilities while preserving the inherent reasoning abilities of MLLMs. For the slow system, we propose a Chain-of-Thought training strategy for failure correction, designed to mimic human reflection after a manipulation failure. Specifically, our model learns to identify the causes of action failures, adaptively seek expert feedback, reflect on the current failure scenario, and iteratively generate corrective actions, step by step. Furthermore, a continuous policy learning method is designed based on successfully corrected samples, enhancing the fast system's adaptability to the current configuration. We compare SC-VLA with the previous SOTA VLA in both simulation and real-world tasks, demonstrating an efficient correction process and improved manipulation accuracy on both seen and unseen tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2405_17418 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | A Self-Correcting Vision-Language-Action Model for Fast and Slow System Manipulation Li, Chenxuan Liu, Jiaming Wang, Guanqun Li, Xiaoqi Chen, Sixiang Heng, Liang Xiong, Chuyan Ge, Jiaxin Zhang, Renrui Zhou, Kaichen Zhang, Shanghang Computer Vision and Pattern Recognition Recently, some studies have integrated Multimodal Large Language Models into robotic manipulation, constructing vision-language-action models (VLAs) to interpret multimodal information and predict SE(3) poses. While VLAs have shown promising progress, they may suffer from failures when faced with novel and complex tasks. To emulate human-like reasoning for more robust manipulation, we propose the self-corrected (SC-)VLA framework, which integrates fast system for directly predicting actions and slow system for reflecting on failed actions within a single VLA policy. For the fast system, we incorporate parameter-efficient fine-tuning to equip the model with pose prediction capabilities while preserving the inherent reasoning abilities of MLLMs. For the slow system, we propose a Chain-of-Thought training strategy for failure correction, designed to mimic human reflection after a manipulation failure. Specifically, our model learns to identify the causes of action failures, adaptively seek expert feedback, reflect on the current failure scenario, and iteratively generate corrective actions, step by step. Furthermore, a continuous policy learning method is designed based on successfully corrected samples, enhancing the fast system's adaptability to the current configuration. We compare SC-VLA with the previous SOTA VLA in both simulation and real-world tasks, demonstrating an efficient correction process and improved manipulation accuracy on both seen and unseen tasks. |
| title | A Self-Correcting Vision-Language-Action Model for Fast and Slow System Manipulation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2405.17418 |