analiza aiqbench
CT-VAM addresses the latency bottleneck in vision-language-action models by decoupling high-level semantic reasoning from low-frequency execution, enabling compact local controllers to run at high frequencies. The architecture achieves competitive success rates with only 68M parameters by routing heterogeneous inputs through a stream-separated decoder that prevents sensory tokens from overwhelming task conditions. This approach offers a practical pathway for deploying robust robotic control on resource-constrained edge hardware while leveraging large models for cloud-side planning.
CT-VAM: A Cerebello-Thalamic-Inspired Vision-Action Model for Efficient Visuomotor Control
arXiv:2606.09572v1 Announce Type: new Abstract: Vision-language-action models have shown strong promise for robot manipulation, yet raw language is primarily needed to specify task intent rather than to be repeatedly processed during high-frequency low-level execution. Motivated by this separation, we propose a cerebello-thalamic-inspired vision-action model (CT-VAM) for efficient task-conditioned visuomotor control. CT-VAM acts as a compact local execution policy that predicts action chunks from dualview visual observations, proprioception, and a lightweight task condition, potentially enabling a practical cloud-edge paradigm in which high-level semantic reasoning can be handled by large models while fast closed-loop control runs on local hardware. To fuse heterogeneous inputs effectively, CT-VAM introduces TARS (Thalamic Action Routing Stream), a stream-separated conditional attention decoder that independently routes action, visual and task streams, preventing dense sensory tokens from overwhelming compact task-relevant conditions. With only 68M parameters, CT-VAM achieves LIBERO success rates competitive with substantially larger VLA models, while reducing inference latency. Together with flow-consistent inpainting for asynchronous chunk execution, CT-VAM supports high-frequency control and demonstrates robust realworld deployment on resource-constrained robotic platforms.
Abstract:Vision-language-action models have shown strong promise for robot manipulation, yet raw language is primarily needed to specify task intent rather than to be repeatedly processed during high-frequency low-level execution. Motivated by this separation, we propose a cerebello-thalamic-inspired vision-action model (CT-VAM) for efficient task-conditioned visuomotor control. CT-VAM acts as a compact local execution policy that predicts action chunks from dualview visual observations, proprioception, and a lightweight task condition, potentially enabling a practical cloud-edge paradigm in which high-level semantic reasoning can be handled by large models while fast closed-loop control runs on local hardware. To fuse heterogeneous inputs effectively, CT-VAM introduces TARS (Thalamic Action Routing Stream), a stream-separated conditional attention decoder that independently routes action, visual and task streams, preventing dense sensory tokens from overwhelming compact task-relevant conditions. With only 68M parameters, CT-VAM achieves LIBERO success rates competitive with substantially larger VLA models, while reducing inference latency. Together with flow-consistent inpainting for asynchronous chunk execution, CT-VAM supports high-frequency control and demonstrates robust realworld deployment on resource-constrained robotic platforms.
Robotics (cs.RO); Artificial Intelligence (cs.AI)
arXiv:2606.09572 [cs.RO]
(or arXiv:2606.09572v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2606.09572 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 262 cuvinte
Subiecte conexe
SlotVLA: Towards Modeling of Object-Relation Representations in Robotic Manipulation
arXiv · 2 oct. 2026
Dyna-2: General Purpose Physical Agents with Hour-Long Autonomy — Jason Ma, Dyna Robotics
Latent Space · 13 oct. 2026
Humanity’s Last Invention — Richard Socher of Recursive
Latent Space · 14 sept. 2026
The Machines that Make the Machines
NVIDIA · 7 oct. 2026