Metraipolis › archive › dc_80375ec1967b

The World Model Planning Accelerator (WMPA): Hardware-Algorithm Co-Design for Embedded AI in Robotics — arXiv:2611.10316

Source: arXiv (en)
Original: https://arxiv.org/abs/2611.10316
License (as archived): CC0-1.0
Archived: 2026-11-10T18:46:11.638Z
SHA-256 of the archived text: 5be59adb6adc122bde51db8001687185798f0e2f31c171a0da96edbb1b865647
Stable URL of this page: https://metraipolis.com/documento/dc_80375ec1967b

How to cite

The World Model Planning Accelerator (WMPA): Hardware-Algorithm Co-Design for Embedded AI in Robotics — arXiv:2611.10316 — arXiv (en), https://arxiv.org/abs/2611.10316 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_80375ec1967b (Metraipolis archive dc_80375ec1967b, sha256 5be59adb6adc122bde51db8001687185798f0e2f31c171a0da96edbb1b865647)

BibTeX

@misc{dc80375ec1967b,
  title = {The World Model Planning Accelerator (WMPA): Hardware-Algorithm Co-Design for Embedded AI in Robotics — arXiv:2611.10316},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.10316},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_80375ec1967b; sha256 5be59adb6adc122bde51db8001687185798f0e2f31c171a0da96edbb1b865647},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.10316. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

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 and that a GPU consumes at least $5.9\times10^{10}$ times the CMOS energy floor. Transient 3D thermal simulations show that this inefficiency is also a thermal one, since sustained back-to-back planning heats a millimeter-scale silicon tile to 556 K even behind an ideal copper heat spreader. Guided by this chain from bound to bottleneck, we propose the World Model Planning Accelerator (WMPA), a chip architecture for embedded world model planning that combines near-memory computation, real-time entropy estimation, variable-precision rollout units, and copy-on-write latent scratchpads. With a design target of 40x lower power, the same tile stays near ambient at 306 K in simulation, pointing toward practical planning on thermally constrained robots.

Tema degli esploratori: calcolo e infrastrutture (elenco arXiv cs.AR, nuovi lavori del giorno).
Categorie arXiv: Hardware Architecture (cs.AR) ; Machine Learning (cs.LG); Robotics (cs.RO).
Pagina del lavoro: https://arxiv.org/abs/2611.10316 — 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 2026-11-10T18:38:22.448Z dalla pagina https://arxiv.org/list/cs.AR/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).