LATIDIA · Robótica

The World Model Planning Accelerator (WMPA): Co-diseño de hardware-algoritmo para IA integrada en robótica

arXiv: 2611.10316v1Tipo de anuncio: Cross Resumen: La IA integrada está pasando de la percepción a la predicción. Los robots ahora usan modelos mundiales $\unicode{x2014}$ simuladores aprendidos que predicen el resultado de las acciones de los candidatos en una comunicación

La noticia

arXiv:2611.10316v1 Announce Type: cross Abstract: Embedded AI is moving from perception to prediction. Robots now use world models$\unicode{x2014}$learned simulators that predict the outcome of candidate actions in a compressed latent space before any is executed. Every predicted step must be computed on board within a tight power and thermal budget, yet the energy these steps fundamentally require, and how far real hardware sits from it, have not been characterized. We derive lower bounds on the arithmetic operations and the CMOS dynamic energy of a latent state transition from information theory and Boolean circuit complexity. Profiling a DreamerV3-style Recurrent State Space Model (RSSM) shows that every rollout operation is memory-bandwidth bound