Metraipolis › archive › dc_fece7c7b98c2

I Act Therefore I Am: When Is JEPA's Action-Conditioning Enough to Learn Causal Mechanisms? — arXiv:2609.31161

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.31161
License (as archived)
CC0-1.0
Archived
2026-10-08T02:13:09.246Z
SHA-256 of the archived text
eade3cf5358bc198bddb6c177c811e3163326a325a8272c6a5242ab563ebb075
Stable URL of this page
https://metraipolis.com/documento/dc_fece7c7b98c2

How to cite

I Act Therefore I Am: When Is JEPA's Action-Conditioning Enough to Learn Causal Mechanisms? — arXiv:2609.31161 — arXiv (en), https://arxiv.org/abs/2609.31161 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fece7c7b98c2 (Metraipolis archive dc_fece7c7b98c2, sha256 eade3cf5358bc198bddb6c177c811e3163326a325a8272c6a5242ab563ebb075)
BibTeX
@misc{dcfece7c7b98c2,
  title = {I Act Therefore I Am: When Is JEPA's Action-Conditioning Enough to Learn Causal Mechanisms? — arXiv:2609.31161},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.31161},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fece7c7b98c2; sha256 eade3cf5358bc198bddb6c177c811e3163326a325a8272c6a5242ab563ebb075},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.31161. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Recent empirical and theoretical advances suggest that joint-embedding predictive architectures (JEPAs) may learn meaningful representations for action-conditioned prediction of future outcomes, thus becoming one of the foundational structures for world models. However, accurate prediction does not, in general, necessarily imply recovery of underlying causal states that give rise to the observed dynamics. This work investigates when and how JEPAs can recover the underlying causal states from observations. We first introduce a latent variable model, in which high-dimensional observations are generated from latent causal states whose dynamics are governed by action-conditioned transition mechanisms. Based on this formulation, we develop a general information-theoretic objective that combines conditional likelihood maximization for learning transition dynamics with entropy maximization for preserving latent state information. We then establish identifiability conditions under which representations learned by this general objective recover the underlying latent causal states up to component-wise invertible transformations and permutation. One key condition for such identifiability is sufficient action-induced variation in the transition mechanisms. Guided by this finding, we instantiate the general objective with an action-modulated Gaussian additive-noise model, yielding action-modulated JEPA (A-JEPA). Experiments on synthetic environments verify the theoretical findings under the identifiability conditions and robustness to moderate violations, while visual benchmarks demonstrate improved state recovery and transfer to unseen transition mechanisms.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-28 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.31161 — 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 2026-10-08T01:43:06.734Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-28?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).