Metraipolis › archive › dc_12fb04daabb2

Complementary Future Token Pruning for World Action Models — arXiv:2610.19957

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.19957
License (as archived): CC0-1.0
Archived: 2026-10-15T21:58:00.068Z
SHA-256 of the archived text: 473648fa7c7838c326d8e54788206c7efa91887842d37e3701607fcbb6528ba2
Stable URL of this page: https://metraipolis.com/documento/dc_12fb04daabb2

How to cite

Complementary Future Token Pruning for World Action Models — arXiv:2610.19957 — arXiv (en), https://arxiv.org/abs/2610.19957 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_12fb04daabb2 (Metraipolis archive dc_12fb04daabb2, sha256 473648fa7c7838c326d8e54788206c7efa91887842d37e3701607fcbb6528ba2)

BibTeX

@misc{dc12fb04daabb2,
  title = {Complementary Future Token Pruning for World Action Models — arXiv:2610.19957},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.19957},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_12fb04daabb2; sha256 473648fa7c7838c326d8e54788206c7efa91887842d37e3701607fcbb6528ba2},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19957. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

World action models (WAMs) jointly predict future visual states and robot actions, but processing dense future-frame tokens incurs substantial inference latency. We propose CF-Prune, a training-free framework for pruning future-frame tokens in pretrained WAMs. Our key insight is that future visual predictions support action generation and therefore need not preserve every visual detail required for high-fidelity video synthesis. CF-Prune combines instruction and action attention to identify task-relevant tokens during the initial denoising forward pass, even when future-frame and action representations remain highly noisy. An entropy-based phase identification mechanism further retains dense computation during precision phases to mitigate pruning-induced performance degradation. Experiments on RoboLab, RoboCasa, and real-world manipulation tasks demonstrate consistent inference speedups. At a 35% pruning ratio, CF-Prune achieves average speedups of 1.19-1.27x while improving task success rates. At a 50% pruning ratio, it achieves 1.29-1.33x average speedups on both simulation benchmarks, with overall success rates within 1.0 percentage point of the baseline.

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.RO, nuovi lavori del giorno).
Categorie arXiv: Robotics (cs.RO).
Pagina del lavoro: https://arxiv.org/abs/2610.19957 — 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 2026-10-15T21:38:44.488Z dalla pagina https://arxiv.org/list/cs.RO/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).