| title | Rethinking VLM Representation for VLA Initialization |
|---|---|
| authors | Weifeng Lin, Siyuan Huang, Hao Li, Tingwei Chen, Ruichuan An, Xinyu Wei, Jianbo Liu, Hongsheng Li |
| date | 2026-05-25 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2605.25802 |
| affiliation | The Chinese University of Hong Kong |
| upvotes | 9 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf code |
| primer | Vision-language-action models |
| abstract | Vision-Language-Action (VLA) models widely adopt pretrained Vision-Language Models (VLMs) as policy backbones, yet it remains unclear what kind of pretrained VLM representation is useful as a VLA initialization. In this paper, we study VLA initialization as a controlled representation-design problem along three axes: capability-level embodied VQA supervision, parameter-update strategy, and robot-data pretraining. Our experiments show that the original pretrained VLM representation is a key source of action performance. However, embodied VQA adaptation does not yield uniform gains: its benefit depends on downstream bottlenecks, and gains from different capability domains are not simply additive. For update strategy, LoRA provides a more reliable initialization than Full Finetune, indicating that overly reshaping the pretrained representation can weaken VLA initialization. Robot-data pretraining further improves VLA initialization, with the strongest variant obtained by staged LoRA-based training. Together, these findings suggest that effective VLM-to-VLA adaptation should inject action-relevant embodied and robot-trajectory signals while preserving the pretrained VLM representation that remains useful for action learning. |
cite
@misc{lin2026rethinking,
title={Rethinking VLM Representation for VLA Initialization},
author={Weifeng Lin and Siyuan Huang and Hao Li and Tingwei Chen and Ruichuan An and Xinyu Wei and Jianbo Liu and Hongsheng Li},
year={2026},
eprint={2605.25802},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2605.25802}
}