Metraipolis › archive › dc_df9e940d88f2

Time-to-Goal: Learned Temporal Metrics for Long-Horizon Planning with JEPA World Models — arXiv:2611.05847

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.05847
License (as archived)
CC0-1.0
Archived
2026-11-05T18:30:18.083Z
SHA-256 of the archived text
4cb846a4ab1cc8794d8d59d050b191aec61f552867aa45c716a74e92861181de
Stable URL of this page
https://metraipolis.com/documento/dc_df9e940d88f2

How to cite

Time-to-Goal: Learned Temporal Metrics for Long-Horizon Planning with JEPA World Models — arXiv:2611.05847 — arXiv (en), https://arxiv.org/abs/2611.05847 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_df9e940d88f2 (Metraipolis archive dc_df9e940d88f2, sha256 4cb846a4ab1cc8794d8d59d050b191aec61f552867aa45c716a74e92861181de)
BibTeX
@misc{dcdf9e940d88f2,
  title = {Time-to-Goal: Learned Temporal Metrics for Long-Horizon Planning with JEPA World Models — arXiv:2611.05847},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.05847},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_df9e940d88f2; sha256 4cb846a4ab1cc8794d8d59d050b191aec61f552867aa45c716a74e92861181de},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.05847. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Joint-Embedding Predictive Architectures (JEPAs) support planning by comparing predicted latent states with an encoded goal. For this comparison to guide temporal progress, states closer in time to the goal should receive lower costs. However, training representations for prediction and distribution regularization does not directly enforce this ordering. We examine this mismatch in flat LeWorldModel (LeWM) by analyzing how its Euclidean goal cost changes with temporal separation. Our analysis identifies conditions under which states at different long offsets receive similar average costs, weakening their temporal distinction. To learn the missing ordering explicitly, we introduce Time-to-Goal (T2G), a lightweight metric trained on latent triples from offline trajectories. A margin ranking loss encourages smaller distances for temporally nearer states, requiring no rewards or policy labels. The learned metric then replaces the Euclidean planning cost while the JEPA encoder and predictor remain frozen. Diagnostics on AntMaze and FourRoom show that T2G preserves stronger temporal contrast and improves near--far ranking at long offsets. Across four navigation and manipulation benchmarks, T2G improves the average planning success rate from 32.7\% (flat LeWM) to 63.0\% and approaches or even surpasses the performance of the hierarchical H-JEPA baselines. These results show that learning temporal ordering for the goal cost can substantially improve latent planning even with a flat world model.

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno).
Categorie arXiv: Robotics (cs.RO) ; Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2611.05847 — 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 2026-11-05T18:19:49.069Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).