3D-VLA: A 3D Vision-Language-Action Generative World Model

Haoyu Zhen, Xiaowen Qiu, Peihao Chen, Jincheng Yang, Xin Yan, Yilun Du, Yining Hong, Chuang Gan
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:61229-61245, 2024.

Abstract

Recent vision-language-action (VLA) models rely on 2D inputs, lacking integration with the broader realm of the 3D physical world. Furthermore, they perform action prediction by learning a direct mapping from perception to action, neglecting the vast dynamics of the world and the relations between actions and dynamics. In contrast, human beings are endowed with world models that depict imagination about future scenarios to plan action accordingly. To this end, we propose 3D-VLA by introducing a new family of embodied foundation models that seamlessly link 3D perception, reasoning, and action through a generative world model. Specifically, 3D-VLA is built on top of a 3D-based large language model (LLM) and a set of action tokens is introduced to engage with the embodied environment. Furthermore, to inject generation abilities into the model, we train the embodied diffusion models and align them into the LLM for predicting the goal image and point cloud. To train our 3D-VLA, we curate a large-scale 3D embodied instruction dataset by extracting vast 3D-related information from existing robotics datasets. Our experiments on held-in datasets demonstrate that 3D-VLA significantly improves the reasoning, multimodality generation and planning capabilities in embodied environments, showcasing its potential in real-world applications.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-zhen24a, title = {3{D}-{VLA}: A 3{D} Vision-Language-Action Generative World Model}, author = {Zhen, Haoyu and Qiu, Xiaowen and Chen, Peihao and Yang, Jincheng and Yan, Xin and Du, Yilun and Hong, Yining and Gan, Chuang}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {61229--61245}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/zhen24a/zhen24a.pdf}, url = {https://proceedings.mlr.press/v235/zhen24a.html}, abstract = {Recent vision-language-action (VLA) models rely on 2D inputs, lacking integration with the broader realm of the 3D physical world. Furthermore, they perform action prediction by learning a direct mapping from perception to action, neglecting the vast dynamics of the world and the relations between actions and dynamics. In contrast, human beings are endowed with world models that depict imagination about future scenarios to plan action accordingly. To this end, we propose 3D-VLA by introducing a new family of embodied foundation models that seamlessly link 3D perception, reasoning, and action through a generative world model. Specifically, 3D-VLA is built on top of a 3D-based large language model (LLM) and a set of action tokens is introduced to engage with the embodied environment. Furthermore, to inject generation abilities into the model, we train the embodied diffusion models and align them into the LLM for predicting the goal image and point cloud. To train our 3D-VLA, we curate a large-scale 3D embodied instruction dataset by extracting vast 3D-related information from existing robotics datasets. Our experiments on held-in datasets demonstrate that 3D-VLA significantly improves the reasoning, multimodality generation and planning capabilities in embodied environments, showcasing its potential in real-world applications.} }
Endnote
%0 Conference Paper %T 3D-VLA: A 3D Vision-Language-Action Generative World Model %A Haoyu Zhen %A Xiaowen Qiu %A Peihao Chen %A Jincheng Yang %A Xin Yan %A Yilun Du %A Yining Hong %A Chuang Gan %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-zhen24a %I PMLR %P 61229--61245 %U https://proceedings.mlr.press/v235/zhen24a.html %V 235 %X Recent vision-language-action (VLA) models rely on 2D inputs, lacking integration with the broader realm of the 3D physical world. Furthermore, they perform action prediction by learning a direct mapping from perception to action, neglecting the vast dynamics of the world and the relations between actions and dynamics. In contrast, human beings are endowed with world models that depict imagination about future scenarios to plan action accordingly. To this end, we propose 3D-VLA by introducing a new family of embodied foundation models that seamlessly link 3D perception, reasoning, and action through a generative world model. Specifically, 3D-VLA is built on top of a 3D-based large language model (LLM) and a set of action tokens is introduced to engage with the embodied environment. Furthermore, to inject generation abilities into the model, we train the embodied diffusion models and align them into the LLM for predicting the goal image and point cloud. To train our 3D-VLA, we curate a large-scale 3D embodied instruction dataset by extracting vast 3D-related information from existing robotics datasets. Our experiments on held-in datasets demonstrate that 3D-VLA significantly improves the reasoning, multimodality generation and planning capabilities in embodied environments, showcasing its potential in real-world applications.
APA
Zhen, H., Qiu, X., Chen, P., Yang, J., Yan, X., Du, Y., Hong, Y. & Gan, C.. (2024). 3D-VLA: A 3D Vision-Language-Action Generative World Model. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:61229-61245 Available from https://proceedings.mlr.press/v235/zhen24a.html.

Related Material