[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fPnHud-jchO35DNdRE70ofOO9-8yMGdcuZxhKZcPYL8s":3},{"code":4,"msg":5,"data":6},200,"操作成功",{"id":7,"title":8,"content":9,"digest":10,"source":10,"coverPath":11,"thumbsCoverPath":12,"isTop":13,"isShow":14,"baseClick":13,"clickCount":15,"createTime":16,"typeId":17,"isNewest":18,"newsInfoTypeRespVo":19,"voiceUrl":22,"voiceSize":23,"taskId":24,"releaseTime":25,"titleEn":26,"contentEn":27,"voiceUrlEn":28,"taskIdEn":29,"voiceSizeEn":30},1210,"中国工程院院士倪光南：AI+机器人将形成新的庞大生态系统","\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">在2025世界机器人大会上，中国工程院院士倪光南发表了题为《“AI+空间计算”让机器看懂世界》的演讲。\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">他表示，“AI+空间计算”是连接物理与数字世界、推动二维向三维交互跃迁的关键桥梁，正重塑人机交互方式。倪院士还强调视觉是智能的基石，通过重构而非生成物理世界，该技术赋能机器人“看懂”环境。他结合制造业智能化升级的实例，系统阐述了如何通过构建“脑-眼-行动协同”的具身智能系统，将机器人提升至L4、L5级智能水平，最终形成强大的“AI+机器人”生态系统，以释放其作为新质生产力的巨大潜能。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px;\">以下为演讲全文：\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">尊敬的各位嘉宾，女士们、先生们，大家好。很高兴来参加2025世界机器人大会，我和大家分享的题目是《“AI+空间计算”让机器看懂世界》。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">我们知道，“AI+空间计算”开启了二维交互向三维交互发展的一个新的方式，是物理世界和数字之间的桥梁，正在重塑人类机器和世界的交互方式，它是推动机器人落地的关键核心技术。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">一、\u003C\u002Fspan>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">“AI+空间计算”：三维交互的新范式\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">历史上交互范式的改变推动了社会的进步，当前我们进入“语音+视频+空间”交互的新范式阶段，正是从二维交互向三维交互发展的重要时期，空间计算在其中起了重要的作用。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">我们看到空间计算对于物理世界来讲，是对物理世界的重构，而不是生成。生成大语言模型目前引领着科技的发展，但是大语言模型不是我们世界的全部，生成式AI不能完全复制复杂的物理世界。空间计算和AI的融合，正以一种全新的方式重构三维物理世界，拓展了大模型通向物理世界的桥梁，促进了物理世界和数字世界之间的融合。所以我们认为，“AI+空间计算”是当前落实人工智能+行动的一项关键核心技术。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">二、视觉：智能进化的基石\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">视觉是智能的起点。在生物进化史中，智能的起点就是物理性，在物理世界来讲，首先是视觉。视觉可以说是智能的起点，是物理空间感知的基础，是大脑与物理世界相互作用的“桥梁”。视觉几乎影响到整个大脑智能的进化，我们整个大脑都参与了来自外界的视觉信息处理。眼睛和智能进化相辅相成。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">图灵奖获得者杨立昆教授有一个说法，就是一个典型大模型，其学习的信息量大概是10的14次方字节，就是后面14个0，相当于互联网上所有公开文本信息量的总和，这对一个人来讲，可能要花几十万年才可能学到。但是一个4岁的小孩，从他睁开眼睛到4岁的过程中，能学到的视频信息量大概也就是10的14次方。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">这说明我们大语言模型训练所用的文本信息是不够的，我们要认识世界、理解世界还需要用到大量的视频信息，为此我们要重视眼睛。视觉是智能的起点，是物理世界控制感知的基础，是大脑和物理世界作用的桥梁。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">三、应用场景：制造业的智能化革命\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">以一个“AI+机器人”工厂的实际工作情况为例，机器人这里有个绿色的小窗口，它看到的世界基本上和人眼看到的大体相当，但是机器人还加入了AR+空间数据进行理解和一些建模的过程。一个机器人由于有了“眼睛”，它就可以在移动环境中快速地观察环境，知道精确的位置。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">以下表为例，我们来看制造业和机器人在不同的工业化时期中的变化。我们把这里分两个时期，一个是传统工业化，一个是新型工业化。德国传统工业化是工业4.0，而新型工业化相当于工业5.0。作为自动制造业的发展来讲，从自动化到智能化的发展，或者像现在“AI+工业制造”的发展，这是两个时期的发展。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002F892e55760f3a4fe5ac443096724b3a53\u002F641(1).webp\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">对于工厂来讲，它的生产模式在过去，我们强调的是刚性化、标准化。那现在我们要融合什么？柔性制造，融合定制化。作为工厂的管理，过去是人被动地去安排作业，现在是工厂的大脑去自主决策。在过去，工业软件是很多一起应用，现在的情况更多是，工业软件会以工作流的方式渗透到整个的生产工作流程之中。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">我们的数字世界、物理世界，在过去相当于是映射的关系，而现在我们要把数字和物理世界融合起来。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">作为机器人来讲，要从自动化的工序变成一个智能化的主体。作为工厂的主体，它的控制方式在过去是实时操作系统，现在需要具身智能系统来进行控制。参照驾驶，过去来讲，相当于L1到L3是自动化的，现在我们希望L4、L5作为智能化的自主移动的机器人。同样我们的操作，以前用按键、键盘、鼠标、屏幕，而我们现在可以用自然语言，人可以和它交互，叫它工作，从人工编程发展到大模型编程，从固定工位发展到多工位的移动工作，同时我们不需要部署调整，可以即插即用，这样达到一个高精度的工作。在过去这要花很长时间，而现在移动工作可以做到非常快速，并且是用很低的代价就可以达到很高的精度。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">总的来说，我们希望人和机器将来在工厂生产线上是协同的、取长补短的。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">按照目前来看，我们认为机器人产业是未来的新兴产业，可能要达到万亿元的规模。在这样一个巨大的机会里面，我们要通过提高机器人产业，发挥它作为新质生产力的作用。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">四、构建具身智能系统：脑、眼与行动的协同\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">机器人的智能一般来说可以分为三大部分：脑、眼和行动系统。这三部分组在一起变成一个具身智能系统，能够全面体现机器人的智慧，发挥它的作用。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">就过去来讲，我们在机器人的“脑”和“眼”的方面投入不够，相对说是一个短板，需要特别予以加强，以此更好体现机器人的效能。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">正像汽车一样，汽车可以分为L1到L5的不同等级，有些组织也对机器人进行了相似的分类。目前的机器人大体上是L1到L3，我们希望通过“眼睛”和“脑”的发展，把机器人整体的智能水平发展到L4、L5以上，这样的机器人才真正体现出它的能力，特别是在移动、自主、高精度上，能够让机器人在提升生产的效率方面发挥作用。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002F9930a3f53afc464b889e88cfce61e1e1\u002F641(2).webp\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">机器人要从“自动化工具”发展到“AI＋机器人”，为了使机器人达到更高等级的智能，需要从“机器人操作系统”，发展到“机器人智能系统”，即融合环境感知、交互决策、运动控制三大具身智能，可操控机器人并作为机器人智能主体。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">那么第一，机器人必须有自己的“大脑”，机器人的“大脑”就是基于大模型的，现在大模型正在引领各方面科技发展和生产的变革。对于机器人操作系统，我们也需要提升到智能系统的水平，进行架构性的变革。机器人有了“大脑”，再加上有了“眼睛”，以及行动的控制，它会在工厂之中发挥重大的作用，按照与人的要求接受指令、接受任务、自主规划，进行投产操作，适应各种场景的需求。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">第二部分是“眼睛”，它是智能的起点。生物进化的历史表明，机器人的智能系统应该突出“眼睛”的作用。目前来看，采用“AI＋空间计算”的方案，即用普通的单目摄像头加上神经网络学习组成，这种方案可以达到比较好的效果，具有很高的性价比，使得机器人拥有良好的、类似人的视觉，拥有适应性、好用、易用、实用这些特点。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">同时我们也知道机器人要和自身本体，比如它的手脚轮子等发生交互，所以我们需要有一个操作系统。现在来讲，我们建议用开源的AGIROS来支撑机器人的行动，这个开源的操作系统是由中国科学院软件所支持的，AGIROS智能机器人操作系统开源社区是中国科学院软件所倡议发起，旨在通过凝聚智能机器人操作系统产学研用各方力量，共同推动智能机器人操作系统技术及生态的发展，全面推进智能机器人领域的开源开放协同创新，为智能机器人产业夯实基础。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">目前有大批的单位以及开发者加入了这个社区，借助开源的力量，AGIROS将在全球机器人业界引领潮流大大增强机器人智能系统的竞争力。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">综合起来，今后机器人不仅是一个操作系统，而是包含脑眼和行动协同的一个具身智能系统。自动化的机器人发展到人工智能时代的“AI＋机器人”，能具有更好的智能，更好的性价比，还能发挥更大的竞争力和作用。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">这里最后我们可以讲，我们的目标是要构建“AI＋机器人”生态系统，历史上有一些大的生态系统，比如PC时代有“Wintel”，到了移动手机时代有“AA”（Android＋ARM），“RV+OSS”（OpenHarmony、openkylin等)，这都是人类智能体世界，而未来“AI+机器人”也将形成庞大的“AI+RV+脑-眼-行动”生态系统。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002F24bf464b323e4929975123a8fd1a41ae\u002F641(3).webp\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">五、小结\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">我把今天的分享归纳为：第一，在当前人工智能引领变革的时代，在国家实施“人工智能＋行动”的大形势下，我们要把自动化的机器人发展到 “AI＋机器人”，更好地发挥它新质生产力的作用；第二，为了达到这个要求，关键在于提升机器人智能水平，我们要用“脑-眼-行动协同”的系统来提高机器人的智能水平，真正让机器人能够看见世界、理解世界、形容世界。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">谢谢大家。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"color: rgb(187, 187, 187);\">【新闻来源】腾讯科技 文｜熊腿腿 \u003C\u002Fspan>\u003Ca href=\"https:\u002F\u002Fnews.qq.com\u002Frain\u002Fa\u002F20250810A0327400\" rel=\"noopener noreferrer\" target=\"_blank\" style=\"color: rgb(187, 187, 187);\">https:\u002F\u002Fnews.qq.com\u002Frain\u002Fa\u002F20250810A0327400\u003C\u002Fa>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">（本网转发此文章，旨在为读者提供更多的信息资讯，所涉内容不构成投资、消费建议。文章事实如有疑问，请与有关方核实，文章观点非本网观点，仅供读者参考。）\u003C\u002Fspan>\u003C\u002Fp>","","https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002F94afe1ea6db24375bc057f9256cc8685\u002FAI领域.jpg","https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002Fthumbs\u002F94afe1ea6db24375bc057f9256cc8685\u002FAI领域.jpg",0,1,214,"2025-08-12 18:19",2,false,{"id":17,"name":20,"enName":21},"芯位视野","Xinwei Vision","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3A55b8c5e4-1841-455c-bd25-0529606d2bae%3A0.wav?Expires=1754998827&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=ODED35c%2FExUviIP0ocM1u4dn37s%3D",17344466,"55b8c5e4-1841-455c-bd25-0529606d2bae","2025-08-12 18:13","Academician Ni Guangnan, Chinese Academy of Engineering: AI+Robotics Will Form a New Vast Ecosystem","\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">At the 2025 World Robotics Conference, Academician Ni Guangnan delivered a speech titled \"AI + Spatial Computing Makes Machines Understand the World.\"\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">He stated that \"AI + spatial computing\" is a key bridge connecting the physical and digital worlds, driving the transition from two-dimensional to three-dimensional interaction, and reshaping human-machine interaction methods. Academician Ni also emphasized that vision is the foundation of intelligence, and through reconstructing rather than generating the physical world, this technology empowers robots to \"understand\" their environment. He systematically explained how to build a \"brain-eye-action collaboration\" embodied intelligent system through examples of intelligent upgrading in manufacturing, enhancing robots to L4 and L5 levels of intelligence, ultimately forming a powerful \"AI + robotics\" ecosystem to unleash its huge potential as a new type of productive force.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px;\">The full text of the speech is as follows:\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Respected guests, ladies and gentlemen, good morning. I am very pleased to attend the 2025 World Robotics Conference, and I would like to share with you the topic \"AI + Spatial Computing Makes Machines Understand the World.\"\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">We know that \"AI + spatial computing\" has opened up a new way for the development of two-dimensional interaction to three-dimensional interaction, serving as a bridge between the physical world and the digital world, and is reshaping the way humans interact with machines and the world. It is a key core technology that promotes the application of robots.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">One,\u003C\u002Fspan>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\"> \"AI + Spatial Computing\": A New Paradigm of Three-Dimensional Interaction\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Changes in the paradigm of interaction have driven social progress throughout history. We are currently entering a new phase of the \"voice + video + space\" interaction paradigm, which is an important period for the transition from two-dimensional to three-dimensional interaction, and spatial computing plays an important role in it.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">We see that spatial computing for the physical world is the reconstruction of the physical world, not generation. Large language models currently lead technological development, but they are not the whole of our world, and generative AI cannot fully replicate the complex physical world. The integration of spatial computing and AI is redefining the three-dimensional physical world in a brand-new way, expanding the bridge for large models to reach the physical world, and promoting the integration between the physical and digital worlds. Therefore, we believe that \"AI + spatial computing\" is a key core technology for implementing artificial intelligence + action at present.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">Two, Vision: The Foundation of Intelligent Evolution\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Vision is the starting point of intelligence. In the history of biological evolution, the starting point of intelligence is physicality. In the physical world, it is first vision. Vision can be considered the starting point of intelligence, the basis of physical space perception, and the \"bridge\" between the brain and the physical world. Vision almost affects the entire evolution of brain intelligence, and our entire brain participates in processing visual information from the outside. Eyes and intelligent evolution complement each other.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Professor Yand Li Kun, recipient of the Turing Award, has a saying that a typical large model's learning information volume is about 10^14 bytes, that is, 14 zeros after, which is equivalent to the total amount of all publicly available text on the Internet. For a person, this may take tens of thousands of years to learn. However, a four-year-old child, from opening his eyes to four years old, can learn video information volume of about 10^14.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">This indicates that the text information used in training our large language models is insufficient. To understand the world, we need to use a lot of video information. Therefore, we should pay attention to eyes. Vision is the starting point of intelligence, the basis of controlling perception in the physical world, and the bridge between the brain and the physical world.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">Three, Application Scenarios: The Intelligent Revolution in Manufacturing\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Taking the actual working conditions of an \"AI + robot\" factory as an example, there is a small green window here, and the world it sees is basically similar to what humans see, but the robot also adds AR + spatial data for understanding and modeling. Because the robot has \"eyes,\" it can quickly observe the environment and know the exact position in a mobile environment.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Taking the following table as an example, we look at the changes in manufacturing and robots in different periods of industrialization. We divide it into two periods: one is traditional industrialization, and the other is new industrialization. Germany's traditional industrialization is Industry 4.0, while new industrialization is equivalent to Industry 5.0. As for the development of automated manufacturing, from automation to intelligent development, or the current development of \"AI + industrial manufacturing,\" these are two periods of development.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002F892e55760f3a4fe5ac443096724b3a53\u002F641(1).webp\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">For factories, in the past, we emphasized rigidity and standardization. Now what do we want to integrate? Flexible manufacturing, integration of customization. As for factory management, in the past, people passively arranged operations, now the factory's brain makes autonomous decisions. In the past, industrial software was applied together, but now industrial software permeates the entire production workflow in the form of work processes.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Our digital world and physical world were, in the past, in a mapping relationship, but now we want to integrate the digital and physical worlds.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">As for robots, they should change from automated processes to intelligent subjects. As the main body of the factory, the control method in the past was real-time operating systems, and now it needs to be controlled by embodied intelligent systems. Referring to driving, in the past, it was from L1 to L3, which was automated, but now we hope that L4 and L5 are intelligent, self-moving robots. Similarly, our operations, in the past, used buttons, keyboards, mice, screens, but now we can use natural language, allowing people to interact with them, telling them to work, from manual programming to large model programming, from fixed positions to multi-position mobile work, and we don't need to deploy and adjust, we can plug and play, achieving high-precision work. This used to take a long time, but now mobile work can be done very quickly and with low cost to achieve high precision.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">In summary, we hope that in the future, people and machines will be collaborative and complementary on the production line of the factory.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">According to the current situation, we believe that the robot industry is a future emerging industry, possibly reaching a scale of trillions of yuan. In such a huge opportunity, we should enhance the robot industry and leverage its role as a new type of productive force.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">Four, Building Embodied Intelligent Systems: Brain, Eye and Action Collaboration\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Generally speaking, the intelligence of robots can be divided into three parts: brain, eye and action system. These three parts together form an embodied intelligent system, which can comprehensively reflect the wisdom of the robot and play its role.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">In the past, we did not invest enough in the \"brain\" and \"eye\" of the robot, which was a shortcoming, and needed to be particularly strengthened to better reflect the effectiveness of the robot.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Just like cars, cars can be divided into different levels from L1 to L5, and some organizations have made similar classifications for robots. Currently, robots are generally at L1 to L3, and we hope that through the development of \"eyes\" and \"brain,\" the overall intelligence level of robots can be developed to above L4 and L5, so that robots can truly demonstrate their capabilities, especially in mobility, autonomy, and high precision, enabling robots to play a role in improving production efficiency.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002F9930a3f53afc464b889e88cfce61e1e1\u002F641(2).webp\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Robots need to develop from \"automated tools\" to \"AI + robots.\" To make robots reach a higher level of intelligence, we need to develop from \"robot operating systems\" to \"robot intelligent systems,\" integrating environmental perception, interactive decision-making, and motion control, which can control robots and serve as the intelligent subject of robots.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Firstly, robots must have their own \"brain,\" and the robot's \"brain\" is based on large models. Large models are currently leading the development of various technological fields and production transformations. For robot operating systems, we also need to upgrade to the level of intelligent systems and carry out architectural reforms. With a \"brain,\" plus \"eyes\" and action control, robots will play a major role in factories, accepting instructions and tasks according to human requirements, planning independently, and performing production operations, adapting to the needs of various scenarios.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">The second part is the \"eyes,\" which is the starting point of intelligence. The history of biological evolution shows that the intelligent system of robots should highlight the role of \"eyes.\" Currently, using the \"AI + spatial computing\" solution, which is composed of a regular monocular camera and neural network learning, can achieve relatively good results, with high cost performance, enabling robots to have good, similar human vision, with characteristics such as adaptability, usability, and practicality.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">At the same time, we know that robots need to interact with their own bodies, such as their hands, feet, wheels, etc., so we need an operating system. At present, we suggest using the open-source AGIROS to support robot actions. This open-source operating system is supported by the Institute of Software, Chinese Academy of Sciences. The AGIROS intelligent robot operating system open source community was initiated by the Institute of Software, Chinese Academy of Sciences, aiming to gather the strength of the research, industry, education, use, and other parties in the intelligent robot operating system, jointly promote the development of intelligent robot operating system technology and ecology, comprehensively promote the open-source, collaborative innovation in the field of intelligent robots, and lay the foundation for the intelligent robot industry.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Currently, a large number of units and developers have joined this community. With the power of open source, AGIROS will greatly enhance the competitiveness of the robot intelligent system and lead the trend in the global robot industry.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">In summary, in the future, robots will not only be an operating system, but also an embodied intelligent system that includes brain, eye, and action collaboration. Automated robots developing to the \"AI + robot\" era of artificial intelligence will have better intelligence, better cost performance, and can play a greater competitive advantage and role.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Finally, we can say that our goal is to build an \"AI + robot\" ecosystem. Historically, there have been some large ecosystems, such as \"Wintel\" in the PC era, and \"AA\" (Android + ARM) in the mobile phone era, \"RV + OSS\" (OpenHarmony, openkylin, etc.), which are all human intelligent entities. In the future, \"AI + robot\" will also form a vast \"AI + RV + brain-eye-action\" ecosystem.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002F24bf464b323e4929975123a8fd1a41ae\u002F641(3).webp\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">Five, Summary\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">I summarize today's sharing as follows: First, in the current era where artificial intelligence leads transformation, under the grand context of the country implementing the \"artificial intelligence + action\" strategy, we should develop automated robots to \"AI + robots\" to better leverage their role as a new type of productive force; second, to achieve this requirement, the key is to improve the intelligence level of robots, we should use the \"brain-eye-action collaboration\" system to enhance the intelligence level of robots, truly enabling robots to see the world, understand the world, and describe the world.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Thank you all.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"color: rgb(187, 187, 187);\">[News Source] Tencent Technology | By Xiong Tuitui \u003C\u002Fspan>\u003Ca href=\"https:\u002F\u002Fnews.qq.com\u002Frain\u002Fa\u002F20250810A0327400\" rel=\"noopener noreferrer\" target=\"_blank\" style=\"color: rgb(187, 187, 187);\">https:\u002F\u002Fnews.qq.com\u002Frain\u002Fa\u002F20250810A0327400\u003C\u002Fa>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">（This article is republished by this site to provide readers with more information and news. The content does not constitute investment or consumption advice. If there are any questions about the facts of the article, please verify with the relevant parties. The views of the article are not the views of this site, and are for reference only.）\u003C\u002Fspan>\u003C\u002Fp>","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3A8e5473ae-44ca-4ed1-8187-02b88912013f%3A0.wav?Expires=1774838498&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=vnJ6BqNMsGMfz2oH3FXKR%2B3oN%2B4%3D","8e5473ae-44ca-4ed1-8187-02b88912013f",17261036]