摘要：准确的交通预测既需要理解场景动态，也需要合成逼真的未来观测结果。近期基于扩散的视频生成模型能够产生视觉上可信的预测，但通常需要昂贵的端到端训练，并且往往将场景理解与图像合成相互耦合。本文提出一种解耦式预测框架，将未来表征学习与视频生成分离。首先，冻结的 V-JEPA 编码器从已观测交通视频中提取预测性潜在表征，在语义潜在空间中捕捉底层场景动态。随后，一个轻量级潜在对齐模块将这些表征投影到冻结的 Cosmos 扩散模块的条件空间中，从而无需重新训练大型生成模型即可实现未来视频合成。通过冻结所有基础模型、仅训练轻量级对齐模块，所提出的框架在保持预测能力的同时，大幅降低了优化复杂度。AI City Challenge 2026 Track 5 基准上的实验结果表明，该方法取得了 75.1297 分，在竞赛中排名第三。这些结果表明，V-JEPA 学得的预测性世界表征能够有效引导下游视频生成，为端到端扩散式预测提供了一种实用且高效的替代方案。

Abstract: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.