ReconVLA: Reconstructive Vision-Language-Action Model as Effective Robot Perceiver

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Song, Wenxuan, Zhou, Ziyang, Zhao, Han, Chen, Jiayi, Ding, Pengxiang, Yan, Haodong, Huang, Yuxin, Tang, Feilong, Wang, Donglin, Li, Haoang
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866908489858678784
author Song, Wenxuan
Zhou, Ziyang
Zhao, Han
Chen, Jiayi
Ding, Pengxiang
Yan, Haodong
Huang, Yuxin
Tang, Feilong
Wang, Donglin
Li, Haoang
author_facet Song, Wenxuan
Zhou, Ziyang
Zhao, Han
Chen, Jiayi
Ding, Pengxiang
Yan, Haodong
Huang, Yuxin
Tang, Feilong
Wang, Donglin
Li, Haoang
contents Recent advances in Vision-Language-Action (VLA) models have enabled robotic agents to integrate multimodal understanding with action execution. However, our empirical analysis reveals that current VLAs struggle to allocate visual attention to target regions. Instead, visual attention is always dispersed. To guide the visual attention grounding on the correct target, we propose ReconVLA, a reconstructive VLA model with an implicit grounding paradigm. Conditioned on the model's visual outputs, a diffusion transformer aims to reconstruct the gaze region of the image, which corresponds to the target manipulated objects. This process prompts the VLA model to learn fine-grained representations and accurately allocate visual attention, thus effectively leveraging task-specific visual information and conducting precise manipulation. Moreover, we curate a large-scale pretraining dataset comprising over 100k trajectories and 2 million data samples from open-source robotic datasets, further boosting the model's generalization in visual reconstruction. Extensive experiments in simulation and the real world demonstrate the superiority of our implicit grounding method, showcasing its capabilities of precise manipulation and generalization. Our project page is https://zionchow.github.io/ReconVLA/.
format Preprint
id arxiv_https___arxiv_org_abs_2508_10333
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ReconVLA: Reconstructive Vision-Language-Action Model as Effective Robot Perceiver
Song, Wenxuan
Zhou, Ziyang
Zhao, Han
Chen, Jiayi
Ding, Pengxiang
Yan, Haodong
Huang, Yuxin
Tang, Feilong
Wang, Donglin
Li, Haoang
Robotics
Computer Vision and Pattern Recognition
Recent advances in Vision-Language-Action (VLA) models have enabled robotic agents to integrate multimodal understanding with action execution. However, our empirical analysis reveals that current VLAs struggle to allocate visual attention to target regions. Instead, visual attention is always dispersed. To guide the visual attention grounding on the correct target, we propose ReconVLA, a reconstructive VLA model with an implicit grounding paradigm. Conditioned on the model's visual outputs, a diffusion transformer aims to reconstruct the gaze region of the image, which corresponds to the target manipulated objects. This process prompts the VLA model to learn fine-grained representations and accurately allocate visual attention, thus effectively leveraging task-specific visual information and conducting precise manipulation. Moreover, we curate a large-scale pretraining dataset comprising over 100k trajectories and 2 million data samples from open-source robotic datasets, further boosting the model's generalization in visual reconstruction. Extensive experiments in simulation and the real world demonstrate the superiority of our implicit grounding method, showcasing its capabilities of precise manipulation and generalization. Our project page is https://zionchow.github.io/ReconVLA/.
title ReconVLA: Reconstructive Vision-Language-Action Model as Effective Robot Perceiver
topic Robotics
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2508.10333