研究人员开发了MoSA，一个旨在弥合模拟与现实世界物理之间差距的新框架。该方法通过学习捕获材料各向异性和异质性的残余应力算子来增强现有的各向同性物理模拟器。MoSA逐步自适应应力并结合运动约束，以提高动力学学习中的数据效率和泛化能力。该框架已在机器人操作任务中成功验证，展现出卓越的准确性和鲁棒性。 AI
影响 通过提高学习动力学的准确性，增强了机器人技术和其他物理系统中从模拟到现实的迁移能力。
排序理由 这是一篇详细介绍用于改进物理模拟的新框架的研究论文。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
完整方法见我们的编辑标准。
arXiv:2605.22597v1 Announce Type: new Abstract: Learning real-world dynamics from visual observations is crucial for various domains. A common strategy is to calibrate simulators by estimating physical parameters, yet accuracy is ultimately bounded by the underlying physical mode…
Learning real-world dynamics from visual observations is crucial for various domains. A common strategy is to calibrate simulators by estimating physical parameters, yet accuracy is ultimately bounded by the underlying physical models, which often assume materials are homogeneous…