2024-08-13
한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina
At present, technology is heading towards an unprecedented "era of great waves"! Embodied intelligence, like a powerful force behind it, is sweeping and reshaping the global landscape of future industries. On the track, both technology giants and emerging forces are facing a common challenge: how to accurately grasp the "new engine" that is enough to cause industrial change in the surging wave of embodied intelligence?
August 8,The "2024 Technology Innovators Conference" hosted by Zhongguancun Zhiyou Research Institute and co-organized by Zhongguancun Dongsheng Science Park and Lide Robotics Platform was grandly opened at the Renaissance Hotel in Zhongguancun Dongsheng Science Park, BeijingThemed "New Engine of Intelligent Transformation of Embodied Intelligence Industry", this conference brings together more than 1,000 industry elites from home and abroad, including top scientists, frontier explorers, cutting-edge entrepreneurs, outstanding entrepreneurs and senior investors, to jointly explore the frontier innovation, scientific and technological achievement transformation, super scenario search and patient capital venture investment of embodied intelligence, so as to ultimately promote the rational prosperity and industrialization and intelligent transformation upgrade of this field!
△Caption: Yang Pengyu, Director of the Innovation and Entrepreneurship Service Department of Beijing Municipal Science and Technology Commission and Zhongguancun Management Committee
Yang Pengyu, Director of the Innovation and Entrepreneurship Service Department of Beijing Municipal Science and Technology Commission and Zhongguancun Administrative Committee, said in his speech: "With the deepening of a new round of scientific and technological revolution and industrial transformation, disruptive scientific and technological innovation results continue to emerge globally, and future industries have become a strategic field that major countries in the world focus on. At present, Beijing is accelerating the construction of an international science and technology innovation center and Zhongguancun's world-leading science and technology park, and has the first-mover advantage in developing future industries. I hope that 'innovators' will inspire more scientific and technological innovation forces in Beijing's scientific and technological innovation ecosystem and open up a new 'road to innovation'!"
△Caption: Chen Lei, deputy director of the Fourth Industrial Promotion Division of the Zhongguancun Science City Management Committee
Chen Lei, deputy director of the Fourth Industrial Promotion Division of the Zhongguancun Science City Management Committee, said in his speech, "Zhongguancun Zhiyou Research Institute is an important part of Haidian District's technology incubation. It focuses on embodied intelligence, advanced manufacturing and other fields, and has discovered and cultivated a number of high-value innovative technology companies with the model of 'achievement transformation + technology incubation + investment acceleration'. This conference focuses on the latest research progress in embodied intelligence and brings together many technological pioneers in the embodied intelligence large-scale models and robotics industries. We hope to use this conference as an opportunity to conduct in-depth exchanges, jointly explore the innovative development path of embodied intelligence, and deepen the application cooperation of embodied intelligence."
At the meeting, the preparatory launching ceremony of the "Forum of 50 Young Scientists" led by Zhongguancun Zhiyou Research Institute was officially held. In the era of embodied intelligence, young scientists and technology innovators are the main force of innovation and an important force for promoting social progress. The "Forum of 50 Young Scientists" aims to build a platform for co-creation, co-construction, mutual trust and win-win, and provide strong support for the construction of the country's modern industrial system. Dozens of well-known scholars and industry leaders from top academic institutions such as Tsinghua University, Peking University, Beijing University of Aeronautics and Astronautics, Harbin Institute of Technology, Shanghai Jiaotong University, Xi'an Jiaotong University, Xi'an University of Technology, Zhejiang University, University of Science and Technology of China, Tongji University, Fudan University, Beijing Institute of Technology, Beijing University of Posts and Telecommunications participated in and witnessed this important moment.
△Caption: The preparatory and launching ceremony of the "50 Young Scientists Forum"
At the same time, the "Zhiyou X Incubation Camp", which is committed to building the first stop for embodied intelligence/robot entrepreneurship, has officially started! This incubation camp is a professional and ecological incubation camp jointly established by Zhongguancun Zhiyou Research Institute and X-MAN, led by Professor Wang Tianmiao, honorary director of the Robotics Institute of Beihang University. By connecting top domestic and foreign universities, senior industry scientists, and industry leaders, it helps early entrepreneurs "integrate" into the top robot founder community. "Zhiyou X Incubation Camp" will help innovators take the first step in entrepreneurship.
△Caption: The opening ceremony of the "Smart Friends X" incubation camp
In the keynote report sharing session, important experts including Huang Tiejun, Chairman of Beijing Zhiyuan Artificial Intelligence Research Institute and Director of Peking University National Key Laboratory of Multimedia Information Processing, Peng Fangyu, Vice Dean of the School of Mechanical Science and Engineering of Huazhong University of Science and Technology and Chairman and General Manager of National Digital Design and Manufacturing Innovation Center, Wang Tianmiao, Honorary Director of Beihang University Robotics Institute and Dean of Zhongguancun Zhiyou Research Institute, and Wang Yu, Founding Dean of the Robotics Institute of the Hong Kong University of Science and Technology and co-founder and chief scientist of Daimeng Robotics, shared their views on "Development trend of AI big models", "Future technology research and judgment of robots", "Intelligent control of robots based on generative AI", "Evolutionary path for high-quality development of embodied intelligence", and "At ease - from embodied intelligence to embodied skills", providing a feast of ideas for the participants.
△Caption: Huang Tiejun, Chairman of Beijing Zhiyuan Artificial Intelligence Research Institute and Director of Peking University's National Key Laboratory of Multimedia Information Processing
"At present, the emergence of intelligence is only 'static integration', waiting for the 'dynamic emergence' ability of brain-like intelligence to realize true AGI." In the keynote report "The Development Trend of AI Big Models", Huang Tiejun, Chairman of Beijing Zhiyuan Artificial Intelligence Research Institute and Director of Peking University's National Key Laboratory of Multimedia Information Processing, believes that in the next three years, big models will expand to general intelligence such as vision, hearing, embodiment, and action, with larger scale and smaller size, and are expected to become the "brain" of embodied intelligence; in the next 10 years, humans will fully enter the era of intelligence, just like the industrial revolution liberated physical strength, and the electric revolution solved the problem of energy circulation. The science and technology, economy, and social ecology with intelligence as the core will be completed; in the next 20 years, behaviorism (reinforcement learning) and brain-like intelligence (radical connectionism) will play a greater role. In 2045, structural brain-like, functional brain-like, and performance super brain will appear, and physical AGI will arrive.
△Caption: Peng Fangyu, Vice Dean of the School of Mechanical Science and Engineering of Huazhong University of Science and Technology, Chairman and General Manager of the National Digital Design and Manufacturing Innovation Center
In the keynote report "Research and Evaluation of Future Robotic Technology", Peng Fangyu, Vice Dean of the School of Mechanical Science and Engineering of Huazhong University of Science and Technology and Chairman and General Manager of the National Digital Design and Manufacturing Innovation Center, said, "One of the important directions of future manufacturing is to achieve ubiquitous manufacturing through ubiquitous robots, ubiquitous sensors, and ubiquitous intelligence. The future trend of manufacturing is human-machine integration, which includes three levels: human-machine space sharing, human-machine skill transfer, and human-machine hybrid intelligence. The latter two levels will break through the level of robot intelligence and make robots 'skilled craftsmen'."
△Caption: Tencent Chief Scientist Zhang Zhengyou
In the keynote report "Intelligent Control of Robots Based on Generative AI", Tencent Chief Scientist Zhang Zhengyou pointed out that "the current embodied intelligence is 'big model + robot', which is artificially forcibly inserting intelligence into the robot. In the future, we expect that intelligence and the body can be organically combined, so that the robot can emerge intelligently in the interaction with the environment."
Professor Zhang Zhengyou systematically proposed the A2G theoretical framework, which is in the order of the first letters of the seven English words ABCDEFG, and comprehensively explains the multi-dimensional requirements for the development of intelligent robots. Specifically: A stands for AI (Artificial Intelligence), emphasizing that robots must have advanced artificial intelligence capabilities, including voice interaction, visual perception, cognitive understanding and autonomous thinking ability, which is the cornerstone of the operation of intelligent robots; B corresponds to Body, focusing on the optimization design of the robot body, pointing out the importance of exploring the best robot form, and humanoid robots are not the only or best choice, and need to be flexibly designed according to the application scenario; C is Control, which refers to the precise control level of the robot.
The above A, B, and C together constitute the basic capability layer of intelligent robots, laying a solid foundation for the in-depth development of robotics technology.
D stands for Developmental Learning, which emphasizes that robots should achieve self-evolution and capability improvement through continuous interaction with the environment to adapt to increasingly complex and changing task requirements; E stands for EQ (Emotional Quotient), which means that robots must have the ability to understand emotions and establish emotional resonance with humans, while promoting humans' understanding and acceptance of robots' emotions and jointly promoting harmonious coexistence between humans and machines; F stands for Flexible Manipulation, which requires robots to be able to accurately and flexibly operate various tools and objects and efficiently assist humans in completing various tasks. This ability is highly dependent on advanced tactile sensor technology; G, as Guardian Angel, symbolizes the robot's guardian mission, emphasizing that robots should be deeply integrated into the intelligent environment, closely connected with family members, other robots, communities and medical institutions, and become reliable guardians of human life and safety.
△Caption: Wang Tianmiao, Honorary Director of the Robotics Institute of Beihang University and President of Zhongguancun Zhiyou Research Institute
Regarding the "Evolutionary Path for High-Quality Development of Embodied Intelligence", Wang Tianmiao, honorary director of the Robotics Institute of Beihang University and dean of Zhongguancun Zhiyou Research Institute, believes that "the sustainable development path of embodied intelligence requires iteration in underlying innovation and application scenarios." There are two paths for innovation and entrepreneurship to cross bubbles and adjustment cycles. One is for small companies with limited resources to start from specific business needs, define products with new technologies, focus on a certain segment, and continuously improve their core competitiveness through the development path of "specialization and innovation". This idea is easier to survive, and they can also cooperate with large companies or even sell to large companies, which will provide application scenarios and financial support. The other is for platform companies with abundant resources. If they achieve breakthroughs in general technologies, they can be widely used in several industries, which can produce revolutionary scientific and technological progress and industrial expansion, continuously raise funds, integrate upstream and downstream resources, and achieve greater economies of scale and synergy effects, thereby occupying a favorable position as a leader or chain leader in market competition, but the mortality rate is also the highest.
△Caption: Wang Yu, founding dean of the Robotics Institute of the Hong Kong University of Science and Technology, co-founder and chief scientist of Daimeng Robotics
Does a robot have to be humanoid? Does it have to have limbs? Wang Yuru, founding dean of the Robotics Institute of the Hong Kong University of Science and Technology, co-founder and chief scientist of Daimeng Robotics, said, "In some industrial scenarios, we see that workers basically sit in fixed positions, and workpieces are (automatically) delivered to workers, so the requirements for "legs" are quite low. What is more important are arms, hands, and eyes. With the cooperation of a pair of dexterous hands, actions such as grasping and moving can be completed, which may not be completely done through 100% vision."
"Ultimately, when we consider every process and every operation in the scene as a model, that is the cerebellum model. I narrowly call it: embodied skills, that is, it has the skills to complete precise operations independently, and it can acquire these skills through certain learning methods."
Outstanding innovators from the field of embodied intelligence brought more practical sharing from the perspectives of core components in continuous evolution, exploration of AI+robots, and exploration of embodied skills in vertical scenarios.
△Caption: Wang Xingxing, founder and CEO of Yushu Technology
In the keynote speech of "Embodied Intelligence is the Most Effective Way to Achieve AGI", Wang Xingxing, founder and CEO of Yushu Technology, believes that "the deployment of physical robots and data collection are the key to embodied intelligence technology leading to AGI." In his opinion, there are four directions in which deep reinforcement learning can do better at present and in the near future, namely: imitating deep reinforcement learning for arbitrary postures or movements such as dance of the whole body; arbitrary and stable movement of the whole body in complex terrain; more complex and dexterous operation of complex objects or parts of the hands; reinforcement learning navigation and obstacle avoidance based on perception data such as images or depth data.
△Caption: Shao Tianlan, founder and CEO of Mecha-Mind
From the perspective of technological exploration and global business practice, Shao Tianlan, founder and CEO of Mecha-Mind, shared that "the mass deployment of intelligent industrial robots still faces three challenges: new technology, high requirements, and long supply chains." However, technologies such as sensors and artificial intelligence are developing at a high speed, and technologies need to be updated faster. Customers have high requirements for product performance and reliability, which requires continuous high-intensity R&D investment. Today, 3D/AI vision and intelligent industrial robot applications are moving from the early market to the mainstream market. In the future, companies must form a positive cycle of leading technology products, leading business, and leading capital. Deepen core technologies, continuously iterate products, and continuously expand applications with partners.
△Caption: Dong Xiaojian, founder and CEO of Beijing Weijing Intelligent Technology Co., Ltd.
Regarding the topic of "Product Practice of Embodied Intelligence", Dong Xiaojian, founder and CEO of Beijing Weijing Intelligent Technology Co., Ltd., believes that "the core of embodied intelligence is the continuous growth of the intelligent body. Treating robots as humans is the basic essence of embodied intelligence."
In his view, the development of embodied intelligence is an evolutionary process from machine to human, and from human to intelligent agent, which emphasizes the combination of continuous learning and practice, especially through the extensive information collection and screening of multiple sensors (such as vision, touch, hearing, etc.), to build the perception and cognitive ability of intelligent agents. This process is not limited to digital sample learning, but also involves the comprehensive processing, understanding and decision-making of information, so that intelligent agents can respond flexibly in the actual environment.
As the problem of aging in society intensifies, the demand for intelligent agents that can assist humans is growing, and embodied intelligence has therefore attracted much attention. The core of embodied intelligence is that intelligent agents can achieve comprehensive improvements in perception, cognition, interaction, motion control, and decision-making capabilities through continuous learning and growth.
△Caption: Fang Hainan, CMO of Beijing Intime Robotics Technology Co., Ltd.
"The core of robot application lies in dexterous operation, and dexterous hands are the key to achieving this ability." In the sharing of "Core Components: The Key Force Driving the Future of Humanoid Robots", Fang Hainan, CMO of Beijing Intime Robotics Technology Co., Ltd., said that in the current rapid development of the robot industry, companies need to find their own positioning, give full play to their respective advantages, and jointly promote technological progress and industrial upgrading. Dexterous hands, as a key component in robot technology, are essential for achieving complex operations and precision operations.
The industry is gradually realizing that by optimizing design and control, even dexterous hands with fewer degrees of freedom can strike a balance between cost-effectiveness and functionality to meet diverse market demands. As technology continues to mature and application scenarios expand, dexterous hands are becoming an important force driving innovation in the robotics industry.
△Caption: Lingxin Qiaoshou CTO Zhou Yong
In the view of Lingxin Smart Hands CTO Zhou Yong, "The next step in embodied intelligence is multimodal perception and interaction algorithms. 'Dexterous operation' in daily life will become a sign of the true arrival of the embodied intelligence era." Lingxin Smart Hands is committed to supporting the next generation of dexterous hands for embodied intelligence. It has unique advantages in three aspects, including: high degrees of freedom, 20 degrees of freedom that can fully map the movements of the human hand in all directions; multi-sensors (position, force, touch, temperature) vision + touch + visual-tactile sensors; and sufficient data combined with landing scenarios (including synthetic data and real collected data).
△Caption: Ma Yuanwei, Chairman of Weiyi Intelligent Manufacturing Technology Committee and Chief Scientist
From AI quality inspection to embodied intelligent industrial robots, Ma Yuanwei, chairman of the Weiyi Intelligent Manufacturing Technology Committee and chief scientist, found that "the primary problem facing robots in the manufacturing industry is line change. Solidly solving the line tangent change of production lines will become a killer application of embodied intelligence."
In his opinion, the current industrial automation industry is undergoing a transformation from traditional non-standard automation to intelligent robotics technology. This transformation emphasizes production flexibility and rapid line change capabilities to adapt to diverse production needs. With the growing trend of personalized and small-batch production, the flexibility and adaptability of robotics technology have become particularly critical.
The key to success in the industry lies not only in technological innovation, but also in the comprehensiveness of solutions and the effectiveness of market services. Technology suppliers need to provide robotic systems that can quickly adapt to new tasks while ensuring efficient and precise execution capabilities.
The development of the embodied intelligence industry also relies on the integration of comprehensive technologies, including perception, planning, control, and machine learning, to enhance the autonomy and intelligence of robots. The integration of these technologies will enable robots to better understand and adapt to complex industrial environments and perform more diverse tasks. At the same time, the long-term development of the industry will require continuous technological innovation and business model exploration to ensure that technological solutions can meet the broad needs of the manufacturing industry and promote the intelligent transformation of the entire industry.
△Caption: Roundtable on "The Evolving Embodied Intelligence: Subversion, Bottlenecks and Technological Ambition"
In the roundtable forum session of "Boiling Embodied Intelligence: Disruption, Bottlenecks and Technological Ambition", Professor Fang Bin of Beijing University of Posts and Telecommunications, as the moderator of the roundtable, had in-depth exchanges and discussions with Professor Liu Huaping of Tsinghua University, Tao Yong, associate professor and doctoral supervisor of the School of Mechanical Engineering of Beihang University, Gao Fei, tenured associate professor of the School of Control Science and Engineering of Zhejiang University, Wang He, assistant professor of Peking University and director of the Peking University-Galaxy Joint Laboratory of General Embodied Intelligence, and Ruan Lecheng, director of the Intelligent Center of Peking University's Human-Machine Fusion Laboratory and researcher.
Six scientists discussed the reasons for the rapid development of embodied intelligence in recent years, believing that this is closely related to breakthroughs in the field of artificial intelligence. They emphasized the importance of interdisciplinary research on embodied intelligence and pointed out that general humanoid robots may be the future development direction. They also discussed the differences between China and the United States in the field of embodied intelligence and how young researchers can enter this field. Experts generally believe that embodied intelligence will become one of the mainstream directions of artificial intelligence in the future and will play an important role in solving social problems such as an aging population and labor shortages. They pointed out that the long-term development of embodied intelligence requires technological breakthroughs, industrial implementation and interdisciplinary integration.
△Caption: The roundtable of "The Era of Embodied Intelligence: The Venture Capital Approach of Patient Capital"
The roundtable forum on "The Era of Embodied Intelligence: The Venture Capital Approach of Patient Capital" was hosted by Wang Bin, President of Yiou, and had in-depth exchanges and discussions with Huang Zihan, Director of the Intelligent Equipment Department of Beijing Economic and Information Technology Bureau, Zhang Chen, Executive Director of Jing Guorui (Head of Beijing Robot Industry Fund), Zhang Ruijun, Professor of Renmin University of China Business School and Founding Partner of Yarui Capital (Head of Zhiyou Scientist Fund), and Jin Tao, Founding Partner of Gaogu Capital, on core issues such as the technical maturity of embodied intelligence, market application prospects, investment risks and returns.
They believe that in the field of embodied intelligence, both "mind" (brain, algorithm) and "handsmanship" (robot body, hardware) are important and need to be developed in a coordinated manner. From an investment perspective, the hardware part is relatively explicit, and it is more difficult to reach a consensus on the algorithm model part. At present, embodied intelligence may be implemented faster in the industrial field, and home service robots still face many challenges. Investors need to pay attention to companies that can play a key role in the industrial chain, as well as companies with innovative capabilities and hard technology content in software and hardware. At the same time, investors also need to consider the impact of the capital market and policy environment on the industry, as well as the long-term development potential of enterprises.
△Caption: "Reshaping the Future Paradigm: Innovation and Challenges of Core Components of Embodied Intelligence" roundtable scene
In the roundtable forum session on "Reshaping the Future Paradigm: Innovation and Challenges of Core Components of Embodied Intelligence", the host, An Ran, Investment Director of Yarui Capital, together with Chen Zhen, founder and CEO of Xiangke Intelligence, Cha Risu, founder and CEO of Huanjun Technology, Bai Guochao, chairman of Zhixing Robotics, Liu Wuyue, founder and CEO of Blue Dot Touch, and Zhao Yunbiao, senior engineer of Wutong Sensing Control and associate researcher of the School of Physics of Peking University, explored the latest progress of core components of embodied intelligence, the technical barriers they face, and the broad blueprint for future development.
Founders and executives of five core component companies shared their observations on the industry's dynamics and changes in the past year. The main points include that embodied intelligence-related hardware is rapidly miniaturizing and standardizing, the combination of software and hardware is becoming increasingly important, robot application scenarios are expanding, and core components and technologies are becoming increasingly segmented.
The discussion also mentioned that humanoid robots have begun to shift from demonstrating basic movement capabilities to actual operational tasks, and the demand for mechanical control has increased. At the same time, flexible sensors have made progress in materials, functions, and algorithms, which have enhanced robots' understanding and interaction capabilities with the environment. These changes reflect that the embodied intelligence industry is moving in a more in-depth and practical direction, gradually entering the "deep water zone".
△Caption: "Searching for Super Scenarios: Where is Embodied Intelligence "Used"? "Roundtable Scene
In the roundtable forum "Looking for Super Scenarios: Where is Embodied Intelligence "Used"? ", Liu Jinchang, a researcher at the High Technology Center of the National Natural Science Foundation of China and a Level 2 Professional and Technical Expert of the Ministry of Science and Technology, served as the moderator, and had a wonderful dialogue and exchange with Zhao Mingguo, a researcher at the Department of Automation of Tsinghua University, director of the Robot Control Laboratory, and chief scientist of Accelerated Evolution, Ji Chao, chief scientist of iFlytek Robotics, Duo Xuan, vice president of Xiaomi Mobile Phone Department and general manager of Xiaomi Robotics Division, and Shi Zhe, chief digital officer of Foxconn Technology Group and CEO of Cloud Intelligence.
Five industry experts had an inspiring exchange of ideas around the core topics of the relationship between embodied intelligence and humanoid robots, how embodied intelligence can reshape traditional industries, and the search for future "super scenarios". They all agreed that humanoid robots, as carriers of embodied intelligence, have shown the advantages of interactive autonomy and flexibility under the strong drive of reinforcement learning technology, and have become the focus of industry attention.
As a platform technology, embodied intelligence is seen as a potential key to solving the problem of high-efficiency production of multiple varieties and small batches in the industrial field due to its strong versatility. Experts emphasize that in order to realize the transition of embodied intelligence technology from the laboratory to the broad market, especially to deploy hundreds of millions of devices in the "super scene", two core conditions must be met: first, the technological achievements must be consumer-oriented (To C), ensuring the close integration of technology with daily life, thereby stimulating huge market demand; second, a comprehensive leap must be achieved in accuracy and performance to ensure that accurate and efficient services can be provided in various complex environments. This is the only way to the "super scene".
In the current wave of science and technology, embodied intelligence, as a bright new star in the field of artificial intelligence, is leading a new round of technological revolution. The "2024 Technology Innovators Conference" is not only a platform for outstanding innovators to display and communicate, but also a collision and integration of ideas. It has injected strong impetus into the rational development and sustainable iterative evolution of embodied intelligence, and has also made important contributions to the transformation, upgrading and prosperous development of my country's industry. With the successful conclusion of the conference, a new chapter of embodied intelligence is slowly unfolding.
The future is here, let us look forward to more long-term demands and practical applications related to "embodied intelligence" emerging!
*This article is reproduced with permission from QuantumBit
Collect and organize the latest news information and make it available for free for everyone to read
