著名人工智能学者、斯坦福大学教授李飞飞创办的World Labs公司近日完成两轮融资，最新一轮融资额达1亿美元，公司估值超10亿美元，跻身独角兽行列。World Labs专注于开发“空间智能”，旨在使机器能够在三维空间中理解和行动。李飞飞认为，这一技术是人工智能发展的关键，其团队正在实验室中训练机器人进行开门、做三明治等任务。
为了理解“空间智能”的概念，李飞飞展示了猫爪推杯子的图片，解释了人类如何在一瞬间评估三维空间中的物体关系并进行预测和行动。除了李飞飞，许多研究团队也在关注3D视觉与机器人的结合，认为这是AI发展中缺乏的关键。
机器之心发现了一份包含80多篇论文的GitHub存储库“Awesome-Robotics-3D”，这些论文涵盖策略学习、预训练、VLM和LLM、表示、模拟等主题，以及两篇综述论文，对3D视觉在机器人操控领域的最新进展进行了全面分析。这些研究对AI在三维空间中的理解和行动能力有着重要的意义，为机器人技术的未来发展提供了参考和指导。
英语如下：
News Title: “Li Feifei Leads New Trend in Robotics 3D Research”
Keywords: Li Feifei, Robotics 3D, Startup Research
News Content: Li Feifei, a renowned scholar in artificial intelligence and a professor at Stanford University, recently completed two rounds of financing for her founded World Labs company. The latest round of financing reached $100 million, with a valuation exceeding $1 billion, putting the company in the elite category of unicorns. World Labs is dedicated to developing “spatial intelligence,” aiming to enable machines to understand and act within three-dimensional spaces. Li Feifei believes that this technology is pivotal to the advancement of artificial intelligence, and her team is currently training robots to perform tasks such as opening doors and making sandwiches in a laboratory setting.
To illustrate the concept of “spatial intelligence,” Li Feifei showed an image of a cat paw pushing a cup, explaining how humans can assess the relationships between objects in a three-dimensional space and predict and act in an instant. Besides Li Feifei, numerous research teams are also focusing on the integration of 3D vision with robotics, recognizing it as a critical missing piece in the development of AI.
AI tech platform Machine之心 discovered a GitHub repository titled “Awesome-Robotics-3D,” which contains over 80 research papers. These papers cover topics such as strategy learning, pre-training, VLM and LLM, representation, simulation, and two review papers that provide a comprehensive analysis of the latest advancements in 3D vision for robotic manipulation. These studies are of great significance for AI’s understanding and action capabilities in three-dimensional spaces, offering references and guidance for the future development of robotics technology.
【来源】https://www.jiqizhixin.com/articles/2024-08-16-3
Views: 2