Metraipolis › archive › dc_c1c6ebefee43

One-Step Next-Latent Prediction Is Not a World Model — arXiv:2609.36227

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.36227
License (as archived)
CC0-1.0
Archived
2026-10-07T17:32:32.307Z
SHA-256 of the archived text
0f51842d63519ceec17a669841e7aa230e2cf65b71de01a825182741bf5eeaa4
Stable URL of this page
https://metraipolis.com/documento/dc_c1c6ebefee43

How to cite

One-Step Next-Latent Prediction Is Not a World Model — arXiv:2609.36227 — arXiv (en), https://arxiv.org/abs/2609.36227 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_c1c6ebefee43 (Metraipolis archive dc_c1c6ebefee43, sha256 0f51842d63519ceec17a669841e7aa230e2cf65b71de01a825182741bf5eeaa4)
BibTeX
@misc{dcc1c6ebefee43,
  title = {One-Step Next-Latent Prediction Is Not a World Model — arXiv:2609.36227},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.36227},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_c1c6ebefee43; sha256 0f51842d63519ceec17a669841e7aa230e2cf65b71de01a825182741bf5eeaa4},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.36227. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Next-latent prediction fits a map from the current embedding to the next one. LeNEPA carries this objective to time series, replacing the stop-gradient of next-embedding prediction with the isotropy penalty of LeJEPA. A world model is a transition kernel that can be rolled out. The one-step regression identifies a conditional mean, and a mean is a kernel only in special cases. For a linear-Gaussian Markov latent, the mean transition and the innovation covariance are fixed by the one-step problem, and the open-loop squared error at horizon $K$ equals the trace of the sum of the pushed-forward innovation covariances. That error grows with $K$ after the one-step fit is exact. If the conditional mean is nonlinear, composing it is not the multi-step conditional mean. If the observation is a non-injective function of a Markov state, a memoryless one-step map does not determine future observations, while a short window can. An isotropy penalty is a function of the embedding marginal, so its partial derivative in the transition weights is zero. On a scalar autoregression with coefficient $0.9$, the one-step mean squared error is $0.998$ and the $16$-step open-loop error is $5.10$. On a hidden rotation, an eight-step window reaches $16$-step error $0.056$, while the current scalar alone reaches $0.778$. Raising the isotropy weight from $0.1$ to $10$ leaves eight-step latent error inside $[0.78,0.85]$ on three seeds.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-30 in cs.LG).
Categorie arXiv: Machine Learning (stat.ML) ; Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.36227 — 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 2026-10-07T16:59:48.479Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).