Artificial intelligence big model, bionic intelligence, unmanned driver, low level technology ...... these hot cutting-edge technology and industrial deep integration will open what kind of imaginative space?On June 2, with the theme of “chasing the new by the quality, intelligent creation of the future”, the 2024 National Conference on New Quality Productivity and Intelligent Industry Development was held in Beijing and Intelligent Industry Development Conference came to an end in Beijing on June 2, under the theme of “Chasing New Quality, Creating the Future”. In this science and technology event organized by the Chinese Society of Automation, Beijing Science and Technology Association, experts and scholars and industry players around the new quality of productivity, and jointly plan the new development of the intelligent industry, the new future, a scientific and technological leap to promote the future of industrial change unfolded.

Explore the frontier technology highlights

“Against the backdrop of economic globalization, informationization and the acceleration of intelligence, automation and artificial intelligence have become an important force in promoting the development of new quality productivity.” Zheng Nanning, academician of the Chinese Academy of Engineering and chairman of the Chinese Society of Automation, pointed out in his speech at the conference.

Participating experts and scholars agreed that the intelligent industry is not only an important direction for future industrial development, but also provides strong technical support for the upgrading and transformation of traditional industries. Strengthening scientific and technological innovation in the field of automation, accelerating the layout and application of automation technology in the intelligent industry, and promoting the in-depth integration of scientific and technological innovation achievements with key industries is an effective way to promote the development of new quality productivity.

In the view of Zheng Qinghua, academician of the Chinese Academy of Engineering and president of Tongji University, the big model has become the pinnacle of the current artificial intelligence, and the reason why the big model is strong is that it relies on big data, big arithmetic and strong algorithms. Big models are bringing subversive changes to various fields: in the field of education, the new model of future education will break through the existing “teacher-student” dichotomy and become a “teacher-machine-student” ternary model; in the field of scientific research, the big model will bring a paradigm shift in scientific research, which will make a large number of people who need professional artificial intelligence to work in the field. In the field of scientific research, the big model brings about a paradigm shift in scientific research, making a large number of scientific experiments that require professional manual labor to be carried out, and the rapid iteration through artificial intelligence ...... AI is also empowering disciplinary innovation, empowering various traditional disciplines and specialties such as civil engineering, architecture, urban planning, transportation, and so on. “I deeply feel that the traditional disciplines old tree spitting new buds, but also for the development of artificial intelligence to provide new scenes and new application platforms, both the development of artificial intelligence itself, but also promote the transformation and upgrading of traditional disciplines.”

In the field of bionic intelligence, Guo Lei, academician of the Chinese Academy of Sciences and professor of Beijing University of Aeronautics and Astronautics, introduces the theoretical and technological research framework of unmanned system bionic intelligent system in the environment of interference confrontation from the theory of estimation and control of multi-source interference system, describes the research progress of the research team in the core technology of the bionic intelligent system - bionic navigation, bionic manipulation and intelligent control systems, etc., and discusses the basic connotation and development direction of bionic intelligent technologies, such as safety, green and immunity, from the perspectives of methodology, system theory and behavior theory. “I study bird-industrial intelligence and fish-industrial intelligence.” Academician Guo Lei said humorously.

Also inspired by the behavioral characteristics of birds and fish, for example, flocks of birds and fish fly and swim densely without collision, Marcos Papageorgiou, professor emeritus of the Technical University of Crete, Greece, proposes a new paradigm of vehicular traffic in the era of intelligent networked and automated vehicles from the principles of lane-line-free traffic and vehicle nudging, as well as optimal path planning for individual vehicles and groups of vehicles, vehicle with queues for signal-free and lane-free urban intersection crossing and other related technologies.

As a typical representative of strategic emerging industries and new quality productivity, low level economy is a new field of competition among major economies around the world, as well as a new engine for China's economic growth and a new track for scientific and technological innovation. Wang Feiyue, Supervisor of Chinese Society of Automation and researcher of Institute of Automation, Chinese Academy of Sciences, analyzed the core elements of low-altitude technology and pointed out the huge potential of low-level technology in promoting industrial upgrading, improving economic efficiency and enhancing social governance potential.

“Low-level economy is China's next opportunity.” According to Wang Feiyue, over the past 40 years of reform and opening up, China's economy has flourished around the four words of “clothing, food, housing and transportation”, and the economic takeoff in the next 40 years will start from “transportation”, from new energy vehicles to intelligent driving, and then to the low-level economy, which can lower logistics costs. The next 40 years of economic takeoff will start from “traveling”, from new energy vehicles to smart driving, to low level economy, to bring down logistics costs.

In the operation of the real economy, logistics cost is an important part of the enterprise cost system. According to the data released by the China Federation of Logistics and Purchasing, the ratio of China's total social logistics costs to GDP in 2023 is 14.4%, significantly higher than that of developed countries such as Europe, America and Japan. Therefore, reducing the level of logistics costs in China has broad potential and great strategic value for promoting economic growth.

Innovative application and accelerated industrial landing

“Artificial Intelligence is the current hotspot around the world, and the most practical AI landing is the big model, and among the many big models, including the language big model, visual big model, etc., the most practical landing is the end-to-end unmanned driving.” Wang Feiyue said in an exclusive interview with the reporter.

In mines, zones, ports and other relatively closed application scenarios, end-to-end driverless technology is rapidly industrialized landing, especially in the field of intelligent mines with the development goal of “reducing manpower, increasing safety, improving efficiency and low carbon”. At the conference, the Institute of Automation of the Chinese Academy of Sciences, China University of Mining and Technology (Beijing), Wu Chuangyuan Intelligent Vehicle Research Institute, and Zhongke Huituo jointly released the YUKON mine model, marking the entry of intelligent mines into the new era of AI, and YUKON mine model has become an important part of the infrastructure of intelligent mines in the era of AI.

Ge Shirong academician, former president of China University of Mining and Technology (Beijing), believes that AI + mining needs to be in-depth innovation, the need to study the intelligent simulation of miners and the expansion of theoretical methods and technologies, to give mining robots the ability to complete the complex environment operations.

Reporters at the conference site to see, with cartoon miner “skin” Huituo“mine treasure” MineCopilot is the first AI assistant for mining practitioners, “He” He can easily understand natural language commands, complete the tasks given to him by the staff, and provide miners with “shoulder-to-shoulder and hand-to-hand” services in the whole process of mining work, and he will become an all-powerful productivity tool for smart mines in the future.

Chen Long, CEO of Huituo, Huituo Mining Treasure is just one of the first products launched on the basis of Yukon mine model, in addition to the generative parallel intelligent data system GenDS, mine end-to-end autopilot model GenAD, as well as mine knowledge model MineGPT, the industry's first four products will help energy mining enterprises to realize the scale of the smart mine factory style. Replication.

"The development of smart mines has entered the Mining 5.0 era, the core of the era is to realize the step from industrial automation to knowledge automation in mines, in which it is necessary to rely on the three core technology pillars, namely, the operating system, scenario engineering, and the mine big model, to complete the four key intelligences of mine intelligentization, including the equipment intelligence, data intelligence, and algorithm intelligence, Process Intelligence." Chen Long revealed that through nearly a decade of development, Zhongke Huituo has made a number of technological breakthroughs in mine operating system and scenario engineering, while the Yukon large model has been learning by using from the beginning of the construction, and every interaction between the user and the “Mine Treasure” can generate higher-quality data for the construction of intelligent mines. Data is the cornerstone of big model training. At present, Zhongke Huituo is constantly expanding the autopilot data at a rate of 240,000 kilometers/day, and has accumulated a total of 80,000 days of mining business logs, accelerating the completion of the layout of the three core technology pillars of the Mining5.0 era, and injecting new productivity into the development of mines.

“By 2025, Yukon mining big model and its infrastructure intelligence system will be fully landed, empowering the construction of more than 40 smart mines across the country, and by 2030, it is expected to guarantee the environment, safe and efficient operation of more than 500 mines around the world.” Chen Long said.

Advice and counseling to discuss the development plan

Focusing on “grasping the trend of the new round of scientific and technological revolution and industrial change, accelerating the cultivation and growth of new productive forces, and shaping new momentum and new advantages for high-quality development”, the conference also set up eight parallel sessions covering intelligent transportation, embodied intelligence, brain-computer intelligence, data elements, enterprise digital transformation, artificial intelligence education, intelligent medicine, and other hot areas, with participants from academia, academia, and the media. Many academicians and experts and scholars from academia and industry shared the latest research results in their respective fields, conducted in-depth discussions on common issues and hotspots, and actively offered advice and suggestions to discuss the development of new quality productivity and intelligent industry.

With regard to the potential risks of artificial intelligence, Academician Zheng Qinghua pointed out that there are several defects in the big model, such as over-reliance on some patterns in the training data, which leads to “illusion”, over-consumption of data and arithmetic power, catastrophic forgetting in the training of new tasks, and weak logical reasoning ability of the black-box model. In this regard, Academician Zheng Qinghua suggests exploring the establishment of China's independent and controllable machine intelligence model, and puts forward the “neural + symbolic” synergistic approach, to create a human brain memory-inspired machine memory intelligence model and other technical routes.

In the field of low level technology and smart economy, Wang Feiyue pointed out the challenges that come along with technological advancement, such as the formulation of regulations, privacy protection and security issues, and called on the government, enterprises and the public to work together and establish a cooperative mechanism, so as to ensure the healthy development of low altitude technology and provide a solid foundation for the construction of a smart economy.

Source: Economic Reference Daily