LATIDIA · Robótica

GPT-6-Astra entiende el tacto y supera la visión: modelos táctiles, de lenguaje y de acción

arXiv: 2610.28457v1Tipo de anuncio: nuevo Resumen: Los modelos multimodales de propósito general se han mostrado prometedores como políticas de acción directa de robots, pero si comprenden la física subyacente a las interacciones del mundo real y pueden aplicarse a

La noticia

arXiv:2610.28457v1 Announce Type: new Abstract: General-purpose multimodal models have shown promise as direct robot action policies, yet whether they grasp the physics underlying real-world interactions and can apply that understanding to guide action remains underexplored. We investigate this question using contact-rich manipulation as a representative testbed, examining tactile understanding as a concrete instance of this broader capability. Contact-rich manipulation requires interpreting tactile feedback and adjusting actions after obstruction. Existing learned tactile policies acquire these capabilities through specialized pretraining and per-task optimization, with substantial data and training costs. We present GPT-as-Tactile-Policy, which summarizes expert demonstrations into task guidance and tactile references for contact assessment and corrective action. GPT-6-Astra combines this context