Published at 10/20/2026, 12:05:41 AM · 44 views
📰 UCSD Unveils ACT3D: Active Exploration in 3D Space via Monocular VLM Interaction, Outperforming Gemini 3.6 Flash
Researchers from UC San Diego and Georgia Tech introduced ACT3D (arXiv:2610.21783), an autonomous agent for active exploration and reasoning in 3D environments.
💬 What's your take? Join the discussion in the comments.
Read the original Chinese version
📰 UCSD 开源 3D 具身探索智能体 ACT3D：纯单目第一人称交互学习，超越 Gemini 3.6 Flash
加利福尼亚大学圣迭戈分校（UCSD）研究团队在 arXiv:2610.21783 中推出 3D 空间主动探索与推理智能体 ACT3D。针对传统 3D 具身模型依赖昂贵稠密多视角图像、多模态视频流或重型点云编码的痛点，该研究完整保留单目视觉语言模型（VLM）的原生单帧第一人称编码，将 3D 环境作为动作与交互观测场。
💬 你怎么看？欢迎在评论区讨论。
#ACT3D #UCSD #Embodied AI
0 Comments