Metraipolis › archive › dc_cb7ec067bf6a

DT-JEPA: Learning to Predict Future Representations for Decision-Making — arXiv:2610.32657

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.32657
License (as archived)
CC0-1.0
Archived
2026-10-26T20:12:30.081Z
SHA-256 of the archived text
b8f830a3e62b08122b0d0dbaf692adee478786c90dc9be98953dd0b2bd6d815f
Stable URL of this page
https://metraipolis.com/documento/dc_cb7ec067bf6a

How to cite

DT-JEPA: Learning to Predict Future Representations for Decision-Making — arXiv:2610.32657 — arXiv (en), https://arxiv.org/abs/2610.32657 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_cb7ec067bf6a (Metraipolis archive dc_cb7ec067bf6a, sha256 b8f830a3e62b08122b0d0dbaf692adee478786c90dc9be98953dd0b2bd6d815f)
BibTeX
@misc{dccb7ec067bf6a,
  title = {DT-JEPA: Learning to Predict Future Representations for Decision-Making — arXiv:2610.32657},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.32657},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_cb7ec067bf6a; sha256 b8f830a3e62b08122b0d0dbaf692adee478786c90dc9be98953dd0b2bd6d815f},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.32657. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Decision Transformer (DT) has demonstrated strong performance in various offline reinforcement learning (RL) tasks by conditioning action generation on expected future returns and historical trajectories. However, treating expected future returns as the condition provides an incomplete description of the future, as trajectories with different state transitions and behaviors may share similar returns. Moreover, expected future returns inherently depend on task-specific reward functions, which makes policy learning reliant on reward-labeled datasets and restricts its scalability to tasks with similar but unseen rewards. Recently, Joint Embedding Predictive Architectures (JEPA) have emerged as a promising paradigm of self-supervised learning-based world models, which learn predictive representations of future dynamics in latent space rather than reconstructing future states in the original space. Inspired by JEPA, we propose DT-JEPA, an offline pretraining and online fine-tuning framework that learns general knowledge from reward-free offline trajectories and achieves excellent adaptation to unseen tasks. Instead of predicting a single deterministic future embedding, we design a novel probabilistic predictor to model the conditional distribution of future representations in the latent space, thereby capturing both predictive information and uncertainty about future dynamics. Theoretically, we show that 1) the JEPA objective seeks to maximize a variational lower bound on the mutual information between historical representations and future trajectories; and 2) predictive future-conditioned policy learning admits a finite-sample performance bound with an O(N-1/4) estimation term. Experiments on D4RL benchmarks demonstrate that DT-JEPA achieves efficient adaptation across varying offline RL tasks, outperforming traditional DT-based and other offline RL methods.

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/2610.32657 — 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 2026-10-26T19:49:03.422Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).