[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fQSfDMyLlZEsXX8lv7q2plrh7VFL2OmrDP9B9lqxs4rM":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},1174,"王兴兴：让机器人真正干活，这个“临界点”亟待突破 ","\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">智能体机器人，或许正在奔向自己的”ChaGPT时刻”，“在十年之内，应该会出现翻天覆地的变化。”宇树科技王兴兴说。\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002Ff505c6b0ef5f4cf09e8dcd1e031fa987\u002F080491c89053435ebb469530d971f3e4.webp\" width=\"510\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp class=\"ql-align-center\">\u003Cspan class=\"ql-lineHeight-1-75\" style=\"color: rgb(187, 187, 187);\">王兴兴在2025 CCF中国开源大会作特邀报告。 澎湃新闻记者 邹佳雯 图\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">2025年8月2日至3日，中国计算机学会（CCF）主办的2025 CCF中国开源大会（CCF ChinaOSC）在上海召开。大会以“蓄势引领、众行致远”为主题，由上海交通大学和CCF开源发展技术委员会承办，上海交通大学校长、中国科学院院士丁奎岭与CCF开源发展技术委员会主任、中国科学院院士王怀民共同担任大会主席。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">依托上海的国际化创新生态和科技产业集聚优势，大会聚焦开源操作系统、开源芯片、开源大模型、开源具身智能等关键方向，邀请来自学术界与产业界的一线专家与开源领军人物，围绕开源技术的前沿进展展开深度交流。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">宇树科技创始人、CEO王兴兴当日现身大会作了特邀报告。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002Fb301884d6b814aa5b097384c6e73675d\u002F674134bbfedb4afbb0ca0c98afc89cf3.webp\" width=\"793\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp class=\"ql-align-center\">\u003Cspan class=\"ql-lineHeight-1-75\" style=\"color: rgb(187, 187, 187);\">2025 CCF中国开源大会现场。 上海交通大学 供图\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">“最近几年人形机器人相关技术进步非常快速，很大的原因是早些年在四足机器人等领域，全球在开源方面做了很多生态的共建，在这个领域，无论是硬件还是软件，都有非常丰富的资源可供参考。”王兴兴坦言，近年来，国内外许多顶尖院校的师生也依托宇树科技的机器人，做了很多算法的开发，发表了优秀论文，又做了很多开源出来，大大推动了整个行业的技术进步，这也被王兴兴视作宇树科技近年来对行业最大的贡献之一，“事实上，目前的AI产业及整个机器人产业，并不是某一家公司或高校单独完成的，很多时候还是依靠大家的共同努力。”\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">在王兴兴的观察中，2025年上半年国内机器人及具身智能产业非常火热。今年上半年中国智能机器人相关领域的整机厂商和零部件厂商的平均业绩增长或在50%以上，整体业绩表现优异。“整个社会的大众的对机器人的热情点燃了，从需求端把这个产业拉动了。”\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">大众认知中的宇树科技机器人，或许是2025年春晚上的那支舞蹈，事实上，宇树科技的机器人们可以做的还有很多，甚至包括打拳击比赛——这考验着机器人的全身运动能力，需要对十几二十个动作做单独的数据采集，再把每个动作丝滑衔接，同时需要让每个动作具备抗干扰能力，这对算法的要求非常高。“理论上，只要动力系统允许，目前我们可以让机器人学习任何动作，但实际上，这还远远不够。”王兴兴介绍，他们现阶段在研究的，是机器人任意动作的生成——不是简单采集数据、训练做动作，而是实时学习、生成任何动作。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">“我相信，一旦机器人能任意生成任何动作，它就能真正干活了。”这里的“干活”就比如，把机器人带到陌生的会场，给它指令，它能实现去给某位嘉宾倒水。王兴兴举例，这也有点类似于图像和视频生成模型，或者语言生成模型，“这在逻辑上是类似的，我们首先要实现能够生成各种内容，某种意义上，这之后才能让它真的干活。”\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">在王兴兴看来，让机器人真的能任意生成动作、能真正地干活，是人形机器人发展的重要方向，也是当今智能体机器人发展亟待突破的“临界点”之一。硬件方面，机器人未来还需要完善细节、降低成本，增加寿命和可靠性等，但决定性的因素还是机器人大模型、具身智能大模型的“不够用”，“大家已经发现了这个方向并在努力，但还没有突破让机器人真正能干活这个‘临界点’。”\u003C\u002Fstrong>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">而如果达成，王兴兴认为，这将是机器人领域的“ChatGPT时刻”。他估计，行业距离这个“时刻”快则一两年，慢则三五年，“在十年之内，应该会出现翻天覆地的变化。”\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">开幕式上，上海交通大学校长丁奎岭致辞。他表示，开源不仅仅是当前全球科技创新的技术风尚，更是推动开放发展、共同发展的重要路径。大学也要走“开源”的发展之路。一方面，打开学科与学校的边界，以目标和任务为导向，跨学科、跨组织地合作，构建一种“只有岗位，没有单位”的大协作机制，将四大战略科技力量联动起来，充分发挥新型举国体制的优势。另一方面，强化人才与科技成果的供给。各个产业的发展特别是战略性新兴产业的发展离不开人才队伍的支撑和基础研究的突破。大学是人才培养与基础研究的主力军，应该为创新驱动发展提供更加充沛的源头活水。交大和华为构建的“目标共识、人事共通、任务共担、成果共享”的合作新范式，为攻关鸿蒙核心技术提供了坚实的基础和广阔的平台。同时交大也在国内率先成立开源鸿蒙技术俱乐部，不仅在核心技术攻关上奋勇当先，更要注重将开源鸿蒙融入到课程体系中，培养面向下一代操作系统的领军人才。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">在特邀报告环节，多位在人工智能与系统软件领域具有深远影响力的资深专家，围绕大模型、开源硬件与智能操作系统等关键技术方向，分享了具有前瞻性的观点与思考，描绘出了开源技术未来发展的路径蓝图。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">中国工程院院士、清华大学计算机系教授郑纬民在《大模型推理服务在系统层面的挑战》报告中，系统梳理了从数据获取、预处理、训练、微调到推理的完整生命周期，并介绍了其团队研发并开源的高效推理系统Mooncake与KTransformers，展示了系统研究在支撑智能时代工作负载中的核心地位。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">中国科学院院士、北京大学教授、上海交通大学人工智能学院首席顾问鄂维南在《Data-centric AI和数据基础设施建设》报告中指出，AI正在经历从“模型中心”向“数据中心”的范式转变，只有构建高质量、高效率的数据基础设施，才能进一步降低人工智能落地的门槛，推进人工智能的规模化落地。\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>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"color: rgb(187, 187, 187);\">【新闻来源】澎湃新闻 记者 邹佳雯 实习生 许文宁 \u003C\u002Fspan>\u003Ca href=\"https:\u002F\u002Fwww.sohu.com\u002Fa\u002F920235635_260616?scm=10001.1104_13-100000-0_922.0-0.0.a2_5X162X1653&spm=smpc.channel_159.block3_218_AB1PKt_1_fd.25.1754299852556bGVqQAe_1104\" rel=\"noopener noreferrer\" target=\"_blank\" style=\"color: rgb(187, 187, 187);\">http:\u002F\u002Fu5a.cn\u002FFEHSV\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\u002F7007da1446c0441ebae69a800988c2aa\u002FAI领域.jpg","https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002Fthumbs\u002F7007da1446c0441ebae69a800988c2aa\u002FAI领域.jpg",0,1,209,"2025-08-04 17:37",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%3A3b61e754-d2b2-4f6a-8fd1-a7379ad695a8%3A0.wav?Expires=1754306679&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=h72xxuObPEP1kG0QTAW79uZNnic%3D",11500598,"3b61e754-d2b2-4f6a-8fd1-a7379ad695a8","2025-08-04 17:32","Wang Xingxing: Let robots really work, this \"critical point\" needs to be broken through.","\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">Intelligent agent robots may be heading towards their own \"ChatGPT moment\", \"In the next ten years, there should be a tremendous change,\" said Wang Xingxing from Yuzhu Technology.\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002Ff505c6b0ef5f4cf09e8dcd1e031fa987\u002F080491c89053435ebb469530d971f3e4.webp\" width=\"510\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp class=\"ql-align-center\">\u003Cspan class=\"ql-lineHeight-1-75\" style=\"color: rgb(187, 187, 187);\">Wang Xingxing gave a special report at the 2025 CCF China Open Source Conference. Pengpai News reporter Zou Jiawen Photo\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">From August 2 to 3, 2025, the 2025 CCF China Open Source Conference (CCF ChinaOSC) hosted by the Chinese Computer Federation (CCF) was held in Shanghai. The conference, with the theme of \"Storing Power for Leadership and Moving Forward Together,\" was co-organized by Shanghai Jiao Tong University and the CCF Open Source Development Technical Committee, with the president of Shanghai Jiao Tong University and academician of the Chinese Academy of Sciences Ding Kuiling and the chairman of the CCF Open Source Development Technical Committee and academician of the Chinese Academy of Sciences Wang Huaimin jointly serving as the chairpersons of the conference.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Relying on Shanghai's international innovation ecosystem and the advantages of concentrated science and technology industries, the conference focused on key directions such as open source operating systems, open source chips, open source large models, and open source embodied intelligence, inviting frontline experts and open source leaders from academia and industry to engage in in-depth exchanges on the cutting-edge developments of open source technologies.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">On that day, Wang Xingxing, founder and CEO of Yuzhu Technology, attended the conference and delivered a special report.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002Fb301884d6b814aa5b097384c6e73675d\u002F674134bbfedb4afbb0ca0c98afc89cf3.webp\" width=\"793\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp class=\"ql-align-center\">\u003Cspan class=\"ql-lineHeight-1-75\" style=\"color: rgb(187, 187, 187);\">The scene of the 2025 CCF China Open Source Conference. Provided by Shanghai Jiao Tong University\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">\"In recent years, the technological progress in humanoid robots has been very rapid, largely due to the fact that in earlier years, a lot of global efforts were made in open source ecosystems in areas such as quadrupedal robots. In this field, whether hardware or software, there are very rich resources available for reference,\" Wang Xingxing admitted. In recent years, many top universities and their teachers and students have also relied on Yuzhu Technology's robots to develop many algorithms, publish excellent papers, and make many open sources, which greatly promoted the technological progress of the entire industry. This has been regarded by Wang Xingxing as one of the greatest contributions of Yuzhu Technology to the industry in recent years. \"In fact, the current AI industry and the entire robot industry are not completed by a single company or university, but often rely on the joint efforts of everyone.\"\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">In Wang Xingxing's observation, the domestic robot and embodied intelligence industry was very hot in the first half of 2025. In the first half of this year, the average performance growth of complete machine manufacturers and parts manufacturers in the field of intelligent robots in China was over 50%, with outstanding overall performance. \"The public's enthusiasm for robots has been ignited, and it has driven the industry from the demand side.\"\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">The public perception of Yuzhu Technology's robots might be the dance on the 2025 Spring Festival Gala, but in fact, Yuzhu Technology's robots can do much more, even including boxing matches - which tests the robot's full-body movement ability, requiring separate data collection for about ten to twenty actions, then smoothly connecting each action, while also needing each action to have anti-interference capability, which requires very high algorithm requirements. \"Theoretically, as long as the power system allows, we can currently let the robot learn any action, but in reality, this is far from enough.\" Wang Xingxing introduced that what they are currently researching is the generation of arbitrary actions by robots - not just collecting data and training to perform actions, but real-time learning and generating any action.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">\"I believe that once robots can generate any action arbitrarily, they can truly work.\" Here, \"working\" means, for example, taking a robot to an unfamiliar venue, giving it instructions, and having it pour water for a guest. Wang Xingxing gave an example, which is somewhat similar to image and video generation models, or language generation models. \"This is logically similar. We first need to achieve the ability to generate various content. In a sense, only after that can we make it truly work.\"\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">In Wang Xingxing's view, enabling robots to generate arbitrary actions and truly work is an important direction for the development of humanoid robots and also one of the critical points that the current intelligent agent robots need to break through. On the hardware side, robots will still need to improve details, reduce costs, increase lifespan and reliability, etc., but the decisive factor is still the insufficient availability of robot large models and embodied intelligence large models. \"Everyone has already discovered this direction and is trying hard, but has not yet broken through the 'critical point' of making robots truly work.\"\u003C\u002Fstrong>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">If this is achieved, Wang Xingxing believes it will be the \"ChatGPT moment\" in the robotics field. He estimates that the industry is approaching this \"moment\" within one or two years, or three to five years at the latest. \"Within ten years, there should be a tremendous change.\"\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">At the opening ceremony, President of Shanghai Jiao Tong University Ding Kuiling delivered a speech. He stated that open source is not only a technical trend of global scientific and technological innovation today, but also an important path to promote open development and common development. Universities should also take the \"open source\" development path. On one hand, they should open up the boundaries of disciplines and schools, and cooperate across disciplines and organizations, guided by goals and tasks, to build a large collaborative mechanism of \"only positions, no units,\" integrating four strategic scientific forces and fully leveraging the advantages of the new national system. On the other hand, they should strengthen the supply of talent and scientific and technological achievements. The development of various industries, especially strategic emerging industries, cannot be separated from the support of talent teams and breakthroughs in basic research. Universities are the main force in talent cultivation and basic research and should provide more abundant sources of momentum for innovation-driven development. The cooperation model between Jiao Da and Huawei, characterized by \"goal consensus, personnel communication, task sharing, and achievement sharing,\" provides a solid foundation and broad platform for overcoming the core technologies of HarmonyOS. At the same time, Jiao Da has also established the first open source HarmonyOS technology club in the country. Not only does it lead in tackling core technologies, but it also focuses on integrating open source HarmonyOS into the curriculum system to cultivate leading talents for the next-generation operating systems.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">In the special report session, several senior experts with far-reaching influence in the fields of artificial intelligence and system software shared forward-looking views and thoughts on key technical directions such as large models, open source hardware, and intelligent operating systems, outlining the blueprint for the future development of open source technologies.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Professor Zheng Weimin, academician of the Chinese Engineering Academy and professor of the Department of Computer Science at Tsinghua University, in his report titled \"Challenges of Large Model Inference Services at the System Level,\" systematically outlined the complete life cycle from data acquisition, preprocessing, training, fine-tuning to inference, and introduced the efficient inference systems Mooncake and KTransformers developed and open-sourced by his team, demonstrating the core role of system research in supporting workloads in the intelligent era.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Academician of the Chinese Academy of Sciences, Professor E Weinan from Peking University and chief consultant of the School of Artificial Intelligence at Shanghai Jiao Tong University, in his report titled \"Data-centric AI and Data Infrastructure Construction,\" pointed out that AI is undergoing a paradigm shift from \"model-centric\" to \"data-centric.\" Only by building high-quality and efficient data infrastructure can we further lower the threshold for AI implementation and promote the large-scale application of AI.\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>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"color: rgb(187, 187, 187);\">[News Source] Pengpai News reporter Zou Jiawen intern Xu Wenning \u003C\u002Fspan>\u003Ca href=\"https:\u002F\u002Fwww.sohu.com\u002Fa\u002F920235635_260616?scm=10001.1104_13-100000-0_922.0-0.0.a2_5X162X1653&spm=smpc.channel_159.block3_218_AB1PKt_1_fd.25.1754299852556bGVqQAe_1104\" rel=\"noopener noreferrer\" target=\"_blank\" style=\"color: rgb(187, 187, 187);\">http:\u002F\u002Fu5a.cn\u002FFEHSV\u003C\u002Fa>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">（This article is reprinted by this website to provide readers with more information and news. The content involved does not constitute investment or consumption advice. If you have 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 website, and are for reference only.)\u003C\u002Fspan>\u003C\u002Fp>","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3A89166ed3-5783-4cc2-98b6-ab4c19a53dd8%3A0.wav?Expires=1774838505&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=6Bs6ncGH6%2F9307ZwBHPFJHSDyBk%3D","89166ed3-5783-4cc2-98b6-ab4c19a53dd8",15492748]