AI开车时该「多想几步」？它学会了看情况
自动驾驶的「世界模型」会预演未来再决定怎么开，但以前不管路况多简单，它都固定预演几步，浪费算力。这篇让模型自己判断：当前场景风险高就继续预演，路况安全就提前停。核心是两样东西：一个「风险估算器」看当前预演片段暴露了多少危险，一个「开关」衡量继续预演能多提升多少规划质量、值不值得花算力。他们还造了一个反事实驾驶数据集——比如记录「如果没刹车会怎样」——来训练这个判断力。结果在 NAVSIM 和 nuScenes 两个基准上，规划效果最好，同时不必要的预演次数大幅减少。它不是你明天就能装进车里的东西，但指向一个方向：让 AI 学会「什么时候该多想，什么时候该行动」，而不是永远傻算。
📄 原文摘要(英文)
World Action Models (WAMs) improve planning by incorporating future world evolution into action generation, yet existing methods allocate a fixed imagination budget to every scene. We propose RISE (Refining Imagination through SElective Rollout), a system-level adaptive imagination framework that makes sequential Roll/Stop decisions according to the expected planning benefit of continued rollout. At each step, a Latent Evaluator estimates the risk revealed by the current prefix and how much planning could improve if imagination continues, while a Rollout Gate weighs this expected benefit against additional computation cost. Since factual driving logs expose only one realized future, we further construct CounterDrive, a counterfactual dataset with diverse outcomes and risk levels, to enrich future dynamics and provide localized risk supervision. Each retained sample undergoes expert verification and annotation of trajectory validity, incident onset, and causal category, providing a reusable resource for safety-critical world-modeling research. Experiments on NAVSIM and nuScenes show that RISE achieves the best overall planning performance while reducing unnecessary rollout, with additional transfer results supporting its plug-in generality across WAM architectures.