LATIDIA · Robótica

La paradoja de la aceleración: repensar la compensación de la calidad de la velocidad de inferencia en las tareas incorporadas

arXiv: 2606.28529v3Announce Type: replace Resumen: Los modelos de cimientos incorporados se han utilizado recientemente para mejorar la generalización del robot y las tasas de éxito de las tareas. Trabajos anteriores aplican la técnica de inferencia eficiente con pérdidas

La noticia

arXiv:2606.28529v3 Announce Type: replace Abstract: Embodied foundation models have recently been widely used to improve robot generalization and task success rates. Previous works apply lossy efficient-inference techniques such as quantization, pruning, and asynchronous inference, accepting small action quality degradation in exchange for lower per-inference computation cost and inter-action latency. However, unlike traditional static ML tasks, embodied tasks involve repeated interaction with the environment, and task-level performance is determined not only by inference cost, but also by closed-loop effects unique to embodied execution, which remain insufficiently characterized in current efficient-inference studies. In this work, we propose TISED (\underline{T}ask-level \underline{I}nference \underline{S}peedup \underline{E}ffect \underline{D}ecomposition), an analytical framework that unifies diverse lossy lightweighting inference