Metraipolis › archive › dc_2d3beb770a72

Knowing When to Stop: Adaptive Action Chunking via Internal Cross-Attention Dynamics in VLAs — arXiv:2609.00908

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.00908
License (as archived)
CC0-1.0
Archived
2026-10-17T07:18:25.109Z
SHA-256 of the archived text
e9b7d83d2416e662ac8eb128fc06802584d29bc7f3f0712f442d19b3e35d191f
Stable URL of this page
https://metraipolis.com/documento/dc_2d3beb770a72

How to cite

Knowing When to Stop: Adaptive Action Chunking via Internal Cross-Attention Dynamics in VLAs — arXiv:2609.00908 — arXiv (en), https://arxiv.org/abs/2609.00908 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2d3beb770a72 (Metraipolis archive dc_2d3beb770a72, sha256 e9b7d83d2416e662ac8eb128fc06802584d29bc7f3f0712f442d19b3e35d191f)
BibTeX
@misc{dc2d3beb770a72,
  title = {Knowing When to Stop: Adaptive Action Chunking via Internal Cross-Attention Dynamics in VLAs — arXiv:2609.00908},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.00908},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2d3beb770a72; sha256 e9b7d83d2416e662ac8eb128fc06802584d29bc7f3f0712f442d19b3e35d191f},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.00908. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Action chunking is a standard execution strategy in modern Vision-Language-Action (VLA) frameworks, but fixed execution horizons impose a trade-off between efficiency and accuracy. Short chunks require frequent inference and may cause oscillatory behavior, whereas long chunks can become misaligned with newly observed states. We address this limitation with an adaptive action chunking approach based on internal cross-attention dynamics in the action expert. We observe that, as the prediction horizon extends, action-to-observation cross-attention becomes increasingly dispersed and its entropy rises toward a plateau. This pattern is associated with higher action prediction error and provides an online signal that the current observation offers limited grounding for further open-loop execution. Based on this observation, we introduce a training-free truncation mechanism that detects sustained high-entropy plateaus and dynamically selects the execution horizon during inference. The method uses attention weights already computed by the policy and introduces negligible additional overhead. Evaluations on $\pi_{0.5}$ and X-VLA across RoboTwin 2.0, LIBERO, and three real-world manipulation tasks show improved average task success over fixed-horizon and adaptive chunking baselines, while preserving efficient closed-loop control. These results show that cross-attention dynamics can provide a practical internal signal for adaptive action execution in VLAs.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-02 in cs.RO).
Categorie arXiv: Robotics (cs.RO).
Pagina del lavoro: https://arxiv.org/abs/2609.00908 — 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 2026-10-17T07:06:32.693Z dalla pagina https://arxiv.org/catchup/cs.RO/2026-09-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).