$M^2$-VLA: Boosting Vision-Language Models for Generalizable Manipulation via Layer Mixture and Meta-Skills
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866913064771649536 |
|---|---|
| author | Xiao, Siyao Zhang, Yuhong Liu, Zhifang Gao, Zihan Zhang, Jingye Choo, Sinwai Zhong, Dake Wang, Mengzhe Lin, Xiao Zhou, Xianfeng Jia, Jia Wang, Haoqian |
| author_facet | Xiao, Siyao Zhang, Yuhong Liu, Zhifang Gao, Zihan Zhang, Jingye Choo, Sinwai Zhong, Dake Wang, Mengzhe Lin, Xiao Zhou, Xianfeng Jia, Jia Wang, Haoqian |
| contents | Current Vision-Language-Action (VLA) models predominantly rely on end-to-end fine-tuning. While effective, this paradigm compromises the inherent generalization capabilities of Vision-Language Models (VLMs) and incurs catastrophic forgetting. To address these limitations, we propose $M^2$-VLA, which demonstrates that a generalized VLM is able to serve as a powerful backbone for robotic manipulation directly. However, it remains a key challenge to bridge the gap between the high-level semantic understanding of VLMs and the precise requirements of robotic control. To overcome this, we introduce the Mixture of Layers (MoL) strategy that selectively extracts task-critical information from dense semantic features. Furthermore, to facilitate efficient trajectory learning under constrained model capacity, we propose a Meta Skill Module (MSM) that integrates strong inductive biases. Extensive experiments in both simulated and real-world environments demonstrate the effectiveness of our approach. Furthermore, generalization and ablation studies validate the architecture's zero-shot capabilities and confirm the contribution of each key component. Our code and pre-trained models will be made publicly available. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_24182 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | $M^2$-VLA: Boosting Vision-Language Models for Generalizable Manipulation via Layer Mixture and Meta-Skills Xiao, Siyao Zhang, Yuhong Liu, Zhifang Gao, Zihan Zhang, Jingye Choo, Sinwai Zhong, Dake Wang, Mengzhe Lin, Xiao Zhou, Xianfeng Jia, Jia Wang, Haoqian Robotics Current Vision-Language-Action (VLA) models predominantly rely on end-to-end fine-tuning. While effective, this paradigm compromises the inherent generalization capabilities of Vision-Language Models (VLMs) and incurs catastrophic forgetting. To address these limitations, we propose $M^2$-VLA, which demonstrates that a generalized VLM is able to serve as a powerful backbone for robotic manipulation directly. However, it remains a key challenge to bridge the gap between the high-level semantic understanding of VLMs and the precise requirements of robotic control. To overcome this, we introduce the Mixture of Layers (MoL) strategy that selectively extracts task-critical information from dense semantic features. Furthermore, to facilitate efficient trajectory learning under constrained model capacity, we propose a Meta Skill Module (MSM) that integrates strong inductive biases. Extensive experiments in both simulated and real-world environments demonstrate the effectiveness of our approach. Furthermore, generalization and ablation studies validate the architecture's zero-shot capabilities and confirm the contribution of each key component. Our code and pre-trained models will be made publicly available. |
| title | $M^2$-VLA: Boosting Vision-Language Models for Generalizable Manipulation via Layer Mixture and Meta-Skills |
| topic | Robotics |
| url | https://arxiv.org/abs/2604.24182 |