Metraipolis › archive › dc_1be7d1ebff81

MaskVLA: Visual Masking Against Trajectory Overfitting of Vision-Language-Action Model — arXiv:2609.23565

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.23565
License (as archived)
CC0-1.0
Archived
2026-10-09T21:59:49.146Z
SHA-256 of the archived text
0a2a026610eae30f42dbc3e4507682bb8475f2e149c440ab606762964344309f
Stable URL of this page
https://metraipolis.com/documento/dc_1be7d1ebff81

How to cite

MaskVLA: Visual Masking Against Trajectory Overfitting of Vision-Language-Action Model — arXiv:2609.23565 — arXiv (en), https://arxiv.org/abs/2609.23565 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_1be7d1ebff81 (Metraipolis archive dc_1be7d1ebff81, sha256 0a2a026610eae30f42dbc3e4507682bb8475f2e149c440ab606762964344309f)
BibTeX
@misc{dc1be7d1ebff81,
  title = {MaskVLA: Visual Masking Against Trajectory Overfitting of Vision-Language-Action Model — arXiv:2609.23565},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.23565},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_1be7d1ebff81; sha256 0a2a026610eae30f42dbc3e4507682bb8475f2e149c440ab606762964344309f},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.23565. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Vision-Language-Action (VLA) models integrate vision-language understanding with executable robot actions, enabling end-to-end learning for robot control. However, our empirical analysis reveals that existing models exhibit severe trajectory overfitting when finetuned on limited datasets. To guide the model in effectively utilizing wrist camera information, we propose MaskVLA, a masking-based fine-tuning strategy. By randomly masking a small portion of the main camera's visual information, the model is guided to autonomously learn more fine-grained, task-relevant, and effective visual features. This process leads to the emergence of robust policies, thereby enhancing the model's capability to tackle complex manipulation tasks and improving its generalization performance. Our method has been comprehensively evaluated on RoboTwin 2.0, achieving an average success rate improvement of 23.2% and 16.8% compared to $\pi_0$ and OpenVLA-OFT, respectively. Furthermore, experiments on real-world ALOHA robots also demonstrate the effectiveness of our approach.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-22 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI); Robotics (cs.RO).
Pagina del lavoro: https://arxiv.org/abs/2609.23565 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-09T21:45:20.926Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-22?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).