导读
英文标题：VFlash: Fast, Accurate Visual Planning via Flash Primitive Decoding arXiv编号：2611.08449 作者：Yu-Cheng Chou, Shihao Wang, Jin Huang 来源：arXiv 原文链接：htt…
【编辑部】这对你意味着什么
该论文提出的 VFlash 方法可将 3D 点运动预测的生成速度提升约 25 倍，并在四个 MolmoSpaces 任务套件中使世界动作模型的平均成功率从 62.6% 提升至 70.2%。对于从事机器人、自动驾驶或具身 AI 研发的在美华人工程师和研究人员，这意味着视觉规划模块的推理延迟可显著降低，有助于在实时系统中部署更复杂的轨迹预测。论文现已发布在 arXiv（编号 2611.08449），可供直接阅读和复现。
英文标题：VFlash: Fast, Accurate Visual Planning via Flash Primitive Decoding
arXiv编号：2611.08449
作者：Yu-Cheng Chou, Shihao Wang, Jin Huang
来源：arXiv
原文链接：https://arxiv.org/abs/2611.08449
中文摘要：
Models for physical AI tasks, such as robotics and autonomous driving, need not reason solely in language. They can use visual plans--points, bounding boxes, and trajectories--to represent spatial goals and anticipate motion. However, autoregressive vision-language models generate these plans one coordinate token at a time, making richer predictions increasingly costly. We introduce VFlash, a method for fast, accurate visual planning through Flash Primitive Decoding. VFlash exploits the known structure of visual primitives to predict their coordinates jointly, using a lightweight drafter that generates an entire primitive in one forward pass. The layout adapts to the primitive's type, dimensionality, and length, supporting compact boxes and long, variable-length trajectories. Crucially, spatial accuracy need not require reproducing an autoregressive model's exact token sequence. VFlash therefore trains on annotated primitives and directly accepts structurally valid predictions without target-token agreement. Across the evaluated 2D and 3D robot trajectory tasks, VFlash consistently reduces prediction errors relative to autoregressive decoding. Preliminary single-rollout measurements indicate approximately 25x faster generation for 3D point motion forecasting. A world action model conditioned on VFlash plans achieves 70.2% average success across four MolmoSpaces task suites, compared with 62.6% for the baseline.
本文由华人空间 AI 论文助手自动抓取翻译，仅供学术交流。
【编辑部】今日讨论
你在实际项目中是否遇到过自回归视觉规划推理太慢、难以满足实时性要求的情况？
