Metraipolis › archive › dc_dfacea5171ad

SPT-Flow: One Integration, Many Futures in Continuous-Time World Models — arXiv:2610.41273

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.41273
License (as archived)
CC0-1.0
Archived
2026-11-03T05:40:46.358Z
SHA-256 of the archived text
5bc336d85c3a7a2648ba78f2eb93c6be3e6b9118516712643635003210c6ea94
Stable URL of this page
https://metraipolis.com/documento/dc_dfacea5171ad

How to cite

SPT-Flow: One Integration, Many Futures in Continuous-Time World Models — arXiv:2610.41273 — arXiv (en), https://arxiv.org/abs/2610.41273 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_dfacea5171ad (Metraipolis archive dc_dfacea5171ad, sha256 5bc336d85c3a7a2648ba78f2eb93c6be3e6b9118516712643635003210c6ea94)
BibTeX
@misc{dcdfacea5171ad,
  title = {SPT-Flow: One Integration, Many Futures in Continuous-Time World Models — arXiv:2610.41273},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.41273},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_dfacea5171ad; sha256 5bc336d85c3a7a2648ba78f2eb93c6be3e6b9118516712643635003210c6ea94},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41273. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Continuous-time world models generate a latent trajectory through a single ODE integration, but adding stochastic conditioning does not ensure that the model uses it. In our teacher-forced baseline, posterior means vary by only about 1% of the prior scale, and matched and swapped latents yield indistinguishable reconstruction PSNR at the reported precision. We study a supervision-shortcut hypothesis: ground-truth intermediate states can reduce the need for trajectory-associated latent information. We introduce SPT-Flow, which samples one latent per trajectory and matches a free-running ODE rollout to demonstration states at multiple readout times. Re-anchoring experiments support the proposed mechanism. On LIBERO, with both models trained for 30k optimizer steps, SPT-Flow improves best-of-10 PSNR over our retrained ODEWorld baseline by 2.0 and 2.4 dB at the mid-trajectory and final frames and reduces mean LPIPS by 23%, with comparable per-sample inference cost. The rollout control recovers most of the accuracy gain; learned latent conditioning adds diversity and best-of-K gains while reducing mean sample accuracy relative to this control. The diagnostics support trajectory-associated latent utilization; semantic mode coverage under matched conditions in the robotic setting remains unverified.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.41273 — 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 2026-11-03T05:26:25.678Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).