Researchers have developed SPEAR-1, a robotic foundation model designed to improve generalization in robot control by integrating 3D spatial reasoning. Unlike previous models trained primarily on 2D image-language tasks, SPEAR-1 enhances a vision-language model with 3D understanding derived from non-robotic data augmented with 3D annotations. This approach allows SPEAR-1 to achieve state-of-the-art performance using significantly fewer robot demonstrations, outperforming models like $\pi_0$-FAST and $\pi_{0.5}$ while requiring 20 times fewer robotic data samples. AI
影响 Enhances robot control generalization by incorporating 3D understanding, potentially reducing the need for extensive robotic data.
排序理由 This is a research paper detailing a new model and methodology for robotic foundation models.
AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →
