Saved in:
Bibliographic Details
Main Authors: Li, Zhuoling, Ren, Liangliang, Yang, Jinrong, Zhao, Yong, Wu, Xiaoyang, Xu, Zhenhua, Bai, Xiang, Zhao, Hengshuang
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2410.07169
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915145105539072
author Li, Zhuoling
Ren, Liangliang
Yang, Jinrong
Zhao, Yong
Wu, Xiaoyang
Xu, Zhenhua
Bai, Xiang
Zhao, Hengshuang
author_facet Li, Zhuoling
Ren, Liangliang
Yang, Jinrong
Zhao, Yong
Wu, Xiaoyang
Xu, Zhenhua
Bai, Xiang
Zhao, Hengshuang
contents The effectiveness of scaling up training data in robotic manipulation is still limited. A primary challenge in manipulation is the tasks are diverse, and the trained policy would be confused if the task targets are not specified clearly. Existing works primarily rely on text instruction to describe targets. However, we reveal that current robotic data cannot train policies to understand text instruction effectively, and vision is much more comprehensible. Therefore, we introduce utilizing vision instruction to specify targets. A straightforward implementation is training a policy to predict the intermediate actions linking the current observation and a future image. Nevertheless, a single future image does not describe the task target in insufficient detail. To handle this problem, we propose to use sparse point flows to provide more detailed information. Extensive tasks are designed based on real and simulated environments to evaluate the effectiveness of our vision instructed pre-training (VIP) method. The results indicate VIP improves the performance on diverse tasks significantly, and the derived policy can complete competitive tasks like ``opening the lid of a tightly sealed bottle''.
format Preprint
id arxiv_https___arxiv_org_abs_2410_07169
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle VIP: Vision Instructed Pre-training for Robotic Manipulation
Li, Zhuoling
Ren, Liangliang
Yang, Jinrong
Zhao, Yong
Wu, Xiaoyang
Xu, Zhenhua
Bai, Xiang
Zhao, Hengshuang
Robotics
The effectiveness of scaling up training data in robotic manipulation is still limited. A primary challenge in manipulation is the tasks are diverse, and the trained policy would be confused if the task targets are not specified clearly. Existing works primarily rely on text instruction to describe targets. However, we reveal that current robotic data cannot train policies to understand text instruction effectively, and vision is much more comprehensible. Therefore, we introduce utilizing vision instruction to specify targets. A straightforward implementation is training a policy to predict the intermediate actions linking the current observation and a future image. Nevertheless, a single future image does not describe the task target in insufficient detail. To handle this problem, we propose to use sparse point flows to provide more detailed information. Extensive tasks are designed based on real and simulated environments to evaluate the effectiveness of our vision instructed pre-training (VIP) method. The results indicate VIP improves the performance on diverse tasks significantly, and the derived policy can complete competitive tasks like ``opening the lid of a tightly sealed bottle''.
title VIP: Vision Instructed Pre-training for Robotic Manipulation
topic Robotics
url https://arxiv.org/abs/2410.07169