Skip to main content
Cornell University
We gratefully acknowledge support from the Simons Foundation, member institutions, and all contributors. Donate
arxiv logo > cs > arXiv:2408.14368v1

Help | Advanced Search

arXiv logo
Cornell University Logo

quick links

  • Login
  • Help Pages
  • About

Computer Science > Robotics

arXiv:2408.14368v1 (cs)
[Submitted on 26 Aug 2024 (this version), latest version 23 Dec 2024 (v2)]

Title:GR-MG: Leveraging Partially Annotated Data via Multi-Modal Goal Conditioned Policy

Authors:Peiyan Li, Hongtao Wu, Yan Huang, Chilam Cheang, Liang Wang, Tao Kong
View a PDF of the paper titled GR-MG: Leveraging Partially Annotated Data via Multi-Modal Goal Conditioned Policy, by Peiyan Li and 5 other authors
View PDF HTML (experimental)
Abstract:The robotics community has consistently aimed to achieve generalizable robot manipulation with flexible natural language instructions. One of the primary challenges is that obtaining robot data fully annotated with both actions and texts is time-consuming and labor-intensive. However, partially annotated data, such as human activity videos without action labels and robot play data without language labels, is much easier to collect. Can we leverage these data to enhance the generalization capability of robots? In this paper, we propose GR-MG, a novel method which supports conditioning on both a language instruction and a goal image. During training, GR-MG samples goal images from trajectories and conditions on both the text and the goal image or solely on the image when text is unavailable. During inference, where only the text is provided, GR-MG generates the goal image via a diffusion-based image-editing model and condition on both the text and the generated image. This approach enables GR-MG to leverage large amounts of partially annotated data while still using language to flexibly specify tasks. To generate accurate goal images, we propose a novel progress-guided goal image generation model which injects task progress information into the generation process, significantly improving the fidelity and the performance. In simulation experiments, GR-MG improves the average number of tasks completed in a row of 5 from 3.35 to 4.04. In real-robot experiments, GR-MG is able to perform 47 different tasks and improves the success rate from 62.5% to 75.0% and 42.4% to 57.6% in simple and generalization settings, respectively. Code and checkpoints will be available at the project page: this https URL.
Comments: 9 pages, 7 figures, letter
Subjects: Robotics (cs.RO); Artificial Intelligence (cs.AI)
Cite as: arXiv:2408.14368 [cs.RO]
  (or arXiv:2408.14368v1 [cs.RO] for this version)
  https://doi.org/10.48550/arXiv.2408.14368
arXiv-issued DOI via DataCite

Submission history

From: Peiyan Li [view email]
[v1] Mon, 26 Aug 2024 15:46:41 UTC (2,426 KB)
[v2] Mon, 23 Dec 2024 14:08:16 UTC (1,349 KB)
Full-text links:

Access Paper:

    View a PDF of the paper titled GR-MG: Leveraging Partially Annotated Data via Multi-Modal Goal Conditioned Policy, by Peiyan Li and 5 other authors
  • View PDF
  • HTML (experimental)
  • Other Formats
view license
Current browse context:
cs.RO
< prev   |   next >
new | recent | 2024-08
Change to browse by:
cs
cs.AI

References & Citations

  • NASA ADS
  • Google Scholar
  • Semantic Scholar
a export BibTeX citation Loading...

BibTeX formatted citation

×
Data provided by:

Bookmark

BibSonomy logo Reddit logo

Bibliographic and Citation Tools

Bibliographic Explorer (What is the Explorer?)
Connected Papers (What is Connected Papers?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)

Code, Data and Media Associated with this Article

alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)

Demos

Replicate (What is Replicate?)
Hugging Face Spaces (What is Spaces?)
TXYZ.AI (What is TXYZ.AI?)

Recommenders and Search Tools

Influence Flower (What are Influence Flowers?)
CORE Recommender (What is CORE?)
  • Author
  • Venue
  • Institution
  • Topic

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)
  • About
  • Help
  • contact arXivClick here to contact arXiv Contact
  • subscribe to arXiv mailingsClick here to subscribe Subscribe
  • Copyright
  • Privacy Policy
  • Web Accessibility Assistance
  • arXiv Operational Status
    Get status notifications via email or slack