Metraipolis › archive › dc_eb0a6fed30d2
JEPA Guided Diffusion: Predictive Vision-Language Conditioning for Generative Traffic Forecasting — arXiv:2609.21379
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.21379
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-10T02:23:11.909Z
- SHA-256 of the archived text
1e108f9d465a3e8403638562474d1e18acfb9c0c61ac5fecdd5270a9fa382a4f- Stable URL of this page
https://metraipolis.com/documento/dc_eb0a6fed30d2
How to cite
JEPA Guided Diffusion: Predictive Vision-Language Conditioning for Generative Traffic Forecasting — arXiv:2609.21379 — arXiv (en), https://arxiv.org/abs/2609.21379 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_eb0a6fed30d2 (Metraipolis archive dc_eb0a6fed30d2, sha256 1e108f9d465a3e8403638562474d1e18acfb9c0c61ac5fecdd5270a9fa382a4f)
BibTeX
@misc{dceb0a6fed30d2,
title = {JEPA Guided Diffusion: Predictive Vision-Language Conditioning for Generative Traffic Forecasting — arXiv:2609.21379},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.21379},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_eb0a6fed30d2; sha256 1e108f9d465a3e8403638562474d1e18acfb9c0c61ac5fecdd5270a9fa382a4f},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.21379. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Accurate traffic forecasting requires both understanding scene dynamics and synthesizing realistic future observations. Recent diffusion-based video generation models produce visually plausible predictions but require expensive end-to-end training and often entangle scene understanding with image synthesis. In this work, we propose a decoupled forecasting framework that separates future representation learning from video generation. A frozen V-JEPA encoder first extracts predictive latent representations from the observed traffic videos, capturing the underlying scene dynamics in a semantic latent space. A lightweight latent alignment module then projects these representations into the conditioning space of a frozen Cosmos diffusion module, enabling future video synthesis without retraining the large generative model. By freezing all foundation models and training only the lightweight alignment module, the proposed framework substantially reduces optimization complexity while preserving forecasting capability. Experimental results on the AI City Challenge 2026 Track 5 benchmark demonstrate that the proposed method achieved a score of 75.1297, ranking third in the competition. These results suggest that predictive world representations learned by V-JEPA can effectively guide downstream video generation, providing a practical and efficient alternative to end-to-end diffusion-based forecasting. Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-21 in cs.CV). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV). Pagina del lavoro: https://arxiv.org/abs/2609.21379 — 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 2026-10-10T02:07:36.043Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-21?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).