Metraipolis › archive › dc_b85d9908013d
JEPA-Anything: Learning Predictive Models across Different Worlds — arXiv:2609.20800
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.20800
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-10T03:48:34.987Z
- SHA-256 of the archived text
9c5dc2326a78cd23f34e13b6e7bc0e4cc73530dfc05de0a577e1297240d1c68d- Stable URL of this page
https://metraipolis.com/documento/dc_b85d9908013d
How to cite
JEPA-Anything: Learning Predictive Models across Different Worlds — arXiv:2609.20800 — arXiv (en), https://arxiv.org/abs/2609.20800 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_b85d9908013d (Metraipolis archive dc_b85d9908013d, sha256 9c5dc2326a78cd23f34e13b6e7bc0e4cc73530dfc05de0a577e1297240d1c68d)
BibTeX
@misc{dcb85d9908013d,
title = {JEPA-Anything: Learning Predictive Models across Different Worlds — arXiv:2609.20800},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.20800},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_b85d9908013d; sha256 9c5dc2326a78cd23f34e13b6e7bc0e4cc73530dfc05de0a577e1297240d1c68d},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.20800. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
World modeling enables intelligence to anticipate consequences, guide interventions, and learn from interaction. Yet predictive models remain domain-specific: can a common learning principle support world modeling across radically different systems? We introduce JEPA-Anything, a domain-agnostic framework based on orthogonal predictive factorization (OPF). Extending joint-embedding predictive architectures, OPF decomposes latent targets into complementary factors, learns them through dedicated pathways, and recombines them within a shared predictive design. We evaluate JEPA-Anything across seven domains: vision, biology, clinical trajectories, control, molecular dynamics, physical fields, and weather. Experiments span representation learning, intervention prediction, out-of-distribution generalization, and long-horizon dynamics, including 10 matched dynamics tasks, forecasting of over 1,000 clinical events, and 100-step molecular rollouts across four systems. Against matched JEPA baselines, JEPA-Anything improves reported metrics on all 10 dynamics tasks and reduces single-intervention prediction error on Interventional Pong by 34.8%. It achieves the lowest one-step and 100-step molecular errors among compared methods in all four systems. Beyond prediction, a factor-nominated biological intervention receives experimental support in cell co-cultures, patient-derived organoids, tumor fragments, and mice; latent orbital modes recover the Keplerian scaling exponent with a fitted slope of -1.4991. These results support a common factorized predictive principle across heterogeneous worlds, connecting world modeling with intervention and experimentally grounded scientific discovery. Code: this https URL Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-18 in cs.CL). Categorie arXiv: Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2609.20800 — 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 2026-10-10T03:24:08.179Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).