SPREAD：通过几何感知扩散实现空间-物理推理
SPREAD: Spatial-Physical REasoning via geometry Aware Diffusion
以带位姿的场景点云作为几何条件、用图 Transformer 联合建模空间与物理关系，并通过可微碰撞/重力/关系约束引导扩散采样，这套「几何感知条件 + 可微物理引导」的组合可直接迁移到具身智能的仿真场景合成与 3D 场景布局生成任务。
原文依据 explicitly conditioning on posed scene point clouds for geometric awareness
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论自动化三维场景生成对虚拟现实、数字内容创作和具身智能等应用至关重要。计算机图形学更看重美观的布局，而视觉与机器人学则要求场景能够反映真实世界的复杂性；由于缺乏大规模无结构训练数据、且空间与物理建模不足，当前的数据驱动方法难以做到这一点。我们提出 SPREAD，一个基于扩散的框架，它通过图 Transformer 联合学习空间与物理关系，并显式地以带位姿的场景点云作为条件以获得几何感知能力。此外，我们的模型集成了针对碰撞避免、关系约束与重力的可微引导，从而在不牺牲关系上下文的前提下保证场景的物理一致性。在 3D-FRONT 和 ProcTHOR 数据集上的实验表明，该方法在空间关系推理与物理指标上取得了最优性能。此外，在物理仿真前后，我们的方法在场景一致性与稳定性上显著优于基线，证明其能够为具身智能体生成可直接用于仿真的环境。
英文原文摘要 1160 字符
Automated 3D scene generation is pivotal for applications spanning virtual reality, digital content creation, and Embodied AI. While computer graphics prioritizes aesthetic layouts, vision and robotics demand scenes that mirror real-world complexity which current data-driven methods struggle to achieve due to limited unstructured training data and insufficient spatial and physical modeling. We propose SPREAD, a diffusion-based framework that jointly learns spatial and physical relationships through a graph transformer, explicitly conditioning on posed scene point clouds for geometric awareness. Moreover, our model integrates differentiable guidance for collision avoidance, relational constraint, and gravity, ensuring physically coherent scenes without sacrificing relational context. Our experiments on 3D-FRONT and ProcTHOR datasets demonstrate state-of-the-art performance in spatial-relational reasoning and physical metrics. Moreover, our method significantly outperforms baselines in scene consistency and stability during pre- and post-physics simulation, proving its capability to generate simulation-ready environments for embodied AI agents.