Metraipolis › archive › dc_2b61dafbcade

AD-E2E-JEPA: A Joint-Embedding Predictive Architecture For End-to-End Autonomous Driving — arXiv:2609.34085

Source

arXiv (en)

Original

https://arxiv.org/abs/2609.34085

License (as archived)

CC0-1.0

Archived

2026-10-07T21:02:06.534Z

SHA-256 of the archived text

0baf1066ba68eff48fc6e4c36a07449d7c5970bee5a7ada1b7d4dce04ac82931

Stable URL of this page

https://metraipolis.com/documento/dc_2b61dafbcade

How to cite

AD-E2E-JEPA: A Joint-Embedding Predictive Architecture For End-to-End Autonomous Driving — arXiv:2609.34085 — arXiv (en), https://arxiv.org/abs/2609.34085 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2b61dafbcade (Metraipolis archive dc_2b61dafbcade, sha256 0baf1066ba68eff48fc6e4c36a07449d7c5970bee5a7ada1b7d4dce04ac82931)

@misc{dc2b61dafbcade,
  title = {AD-E2E-JEPA: A Joint-Embedding Predictive Architecture For End-to-End Autonomous Driving — arXiv:2609.34085},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.34085},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2b61dafbcade; sha256 0baf1066ba68eff48fc6e4c36a07449d7c5970bee5a7ada1b7d4dce04ac82931},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.34085. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Autonomous driving requires \textit{world models} that can understand the physical world, reason and plan, and operate safely. In this paper, we first systematically evaluate existing action-conditioned joint-embedding predictive architecture (JEPA) world models, including LeWM, DINO-WM, and JEPA-WM for end-to-end autonomous driving (E2EAD). To isolate world-model quality from policy learning, we employ a goal-conditioned zero-shot planning setting that evaluates these models using ground-truth future observations as goals, without training any driving policy. We find that existing JEPA-based world models are either accurate for driving but computationally expensive, or computationally efficient but insufficient for planning. To address this trade-off, we propose \textbf{AD-E2E-JEPA}, which introduces a SIGReg-regularized learnable projector applied to projected patch embeddings. The projector reduces the number of planning patches by $16\times$ and the embedding dimension by $4\times$, achieving a $100\times$ inference speedup while retaining planning performance, with a 0.8-second runtime for an 8-frame rollout over 256 candidate trajectories. \textit{Without} training any driving policy, the world model itself reaches the goals located 20 meters away on average within the displacement of respectively 4.0/2.8 meters, using world-model rollouts over trajectory vocabularies of respectively 256/8,192 candidates. On the NAVSIMv2 benchmark, it achieves 67.3/72.9 EPDMS with multiplicative safety metrics and 84.1/86.5 EPDMS$^{\dagger}$ without them in goal-conditioned zero-shot planning. Experiments further show that the self-supervised pretrained projector improves downstream imitation learning performance from 80.2 to 85.4 EPDMS. The source code is available at this https URL

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI).
Categorie arXiv: Robotics (cs.RO) ; Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.34085 — 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 2026-10-07T20:29:02.849Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).