Metraipolis › archive › dc_48eccecfeb94

DS-VLA: A Dendritic-inspired Vision-Language-Action Model for Robust Action Control — arXiv:2609.32253

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.32253
License (as archived): CC0-1.0
Archived: 2026-10-07T21:00:00.303Z
SHA-256 of the archived text: 651fc5966d46fbb2333b39231e0b0b89f8fed897b23ee3fabef8ac131cc3d54c
Stable URL of this page: https://metraipolis.com/documento/dc_48eccecfeb94

How to cite

DS-VLA: A Dendritic-inspired Vision-Language-Action Model for Robust Action Control — arXiv:2609.32253 — arXiv (en), https://arxiv.org/abs/2609.32253 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_48eccecfeb94 (Metraipolis archive dc_48eccecfeb94, sha256 651fc5966d46fbb2333b39231e0b0b89f8fed897b23ee3fabef8ac131cc3d54c)

BibTeX

@misc{dc48eccecfeb94,
  title = {DS-VLA: A Dendritic-inspired Vision-Language-Action Model for Robust Action Control — arXiv:2609.32253},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.32253},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_48eccecfeb94; sha256 651fc5966d46fbb2333b39231e0b0b89f8fed897b23ee3fabef8ac131cc3d54c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.32253. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Vision-language-action (VLA) models have achieved strong performance in language-conditioned manipulation, yet success under nominal evaluation does not necessarily translate into robust closed-loop behavior when executed actions are transiently corrupted. We introduce DS-VLA, a dendritic-inspired action architecture that incorporates dendritic spiking dynamics into VLA control to address this limitation. Specifically, to enable modularized feature processing and temporal information integration, DS-VLA equips action neurons with multiple sparsely connected dendritic branches, each featuring heterogeneous, learned decay factors. Furthermore, to suppress unreliable state updates while preserving task-relevant historical information, we introduce a neuron-wise inhibitory gate that adaptively regulates the admission of new multimodal evidence into dendritic states prior to somatic dynamics. We evaluate DS-VLA on all four LIBERO suites under both nominal rollouts and a unified closed-loop action-perturbation protocol. DS-VLA achieves a 91.6\% average nominal success rate and an 87.35\% average perturbed success rate, retaining 95.4\% of its nominal performance. Under the same reported perturbation setting, OpenVLA-OFT, FAST, $\pi_0$, and GR00T achieve 39.45\%, 23.90\%, 28.55\%, and 30.75\%, respectively. A controlled ablation isolates the contribution of neuron-wise shared inhibition, while analyses of neural dynamics and post-perturbation trajectories associate robust performance with selective evidence suppression and effective behavioral recovery. Together, these results demonstrate that integrating brain-inspired computational mechanisms offers a promising architectural prior for robust embodied intelligence beyond merely scaling vision-language backbones or generative action decoders.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI).
Categorie arXiv: Robotics (cs.RO) ; Artificial Intelligence (cs.AI); Neural and Evolutionary Computing (cs.NE).
Pagina del lavoro: https://arxiv.org/abs/2609.32253 — 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 2026-10-07T20:29:02.269Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).