A 2-meter×1-meter integral gas radiant heating belt product is placed in the 2023 Zhongguancun Forum Information Technology and Intelligent Manufacturing Exhibition Area, which is the latest "hydrogen-doped 40% combustion system" of Calioma, Italy. In addition to the big guy "Amway", Yang Siping, president of Calioma China in Italy, who always deals with traditional industries, also spent a lot of time introducing intelligent energy-saving solutions to visitors during the exhibition.
"The smart heat supply management platform collects heat supply data in real time, unified access and storage of heating equipment information, remote management of heat supply, intelligent control, and a factory uses a smart energy management system combined with tall space radiation heating technology to save energy and reduce carbon emissions by more than 50%." The platform mentioned by Yang Siping is the smart heat management system solution cooperated by Kallioma and the Chinese Academy of Sciences Software Institute.
Energy is only one of the N combinations of "AI+", and the intelligent exploration of various industries can be seen everywhere at this Zhongguancun Forum exhibition. According to the China Academy of Information and Communications Technology, the scale of the mainland's core artificial intelligence industry reached 508 billion yuan last year, a year-on-year increase of 18%. The grassroots computing power platform and open source technology system support the integration of generative AI and basic discriminative AI, making the cognition that artificial intelligence is equivalent to the fourth industrial revolution more certain.
Focus on core technologies
In 2021-2022, when ChatGPT was not yet popular, Beijing KLCII Artificial Intelligence Research Institute (hereinafter referred to as "KLCII") predicted more than once that the era of artificial intelligence big models was coming. At that time, KLCII's large-model projects "Wudao" 1.0 and 2.0, the intelligent computing platform Jiuding, and the large-model technology open-source system FlagOpen are the basis for promoting artificial intelligence into a new process.
At the beginning of this year, KLCII and more than 30 industry-university-research units jointly undertaken the flagship project "Artificial Intelligence Basic Model Support Platform and Evaluation Technology", and after two months of intensive research, developed the FlagOpen large model technology open source system, including large model algorithms, models, data, tools, evaluation and other important components. Based on FlagOpen, domestic and foreign developers can quickly start various large model attempts, development and research work. The FlagOpen large model basic software open source system is gradually realizing complete support for a variety of deep learning frameworks and AI chips.
The emphasis on artificial intelligence at the national level is also progressing layer by layer. In the Outline of the 14th Five-Year Plan for National Economic and Social Development of the People's Republic of China and the Long-Range Goals for 2035 (hereinafter referred to as the "Planning Outline") released in 2021, there are 57 expressions related to "intelligence" and "wisdom". Among the key research directions in the frontier fields of science and technology listed in the "Planning Outline", the first is a new generation of artificial intelligence: cutting-edge basic theoretical breakthroughs, special chip research and development, deep learning framework and other open source algorithm platform construction, learning reasoning and decision-making, image graphics, voice and video, natural language recognition processing, etc.
The highest award in the field of intelligent science and technology in mainland China announced in early May this year, "Wu Wenjun Artificial Intelligence Science and Technology Award", which focuses on the above fields, a total of 60 award-winning project achievements were selected, and the two awards of the Technology Invention Award and the Science and Technology Progress Award also selected the special prize for the first time, respectively, won by the Tao Jianhua team of the Institute of Automation of the Chinese Academy of Sciences and the team of Wang Haifeng, the chief technology officer of Baidu, Tao Jianhua's team completed "key technologies for audio character generation and identification", and Wang Haifeng led the team to complete General Dialogue Technology and Application of Knowledge and Deep Learning Fusion". Zheng Nanning, an academician of the Chinese Academy of Engineering, has achieved a series of important achievements in the fields of computer vision application, basic theory, artificial intelligence chips, and visual intelligent systems, and won the "Wu Wenjun Highest Achievement Award in Artificial Intelligence".
The synergy of partners
It has become a consensus that the development of artificial intelligence should take the integrated application of industries and the digital transformation of industries as the core goals, as can be seen from the recent launch of the "Beijing General Artificial Intelligence Industry Innovation Partnership Program" by the Beijing Municipal Bureau of Economy and Information Technology, the Beijing Municipal Science and Technology Commission, the Zhongguancun Administrative Committee, and the Municipal Development and Reform Commission. Among them, the first batch of 39 member units are divided into five categories: computing power partners, data partners, model partners, application partners and investment partners, including Alibaba Cloud and Beijing Super Cloud Computing Center. 9 data partners such as Beijing Big Data Center and Zhihu; Baidu, Ali Damo Academy and other 7 model partners; There are 13 application partners including Beijing SME Service Center and Haidian City Brain and 8 investment partners including IDG. The goal of the partnership plan is to basically form a general artificial intelligence industry development pattern with complete elements, leading technology, complete ecology, and strong support for the high-quality development of the digital economy by 2025.
What is the meaning of partnering? KLCII has a voice.
In 2018, KLCII, an innovative R&D institution in the field of AI, was established to promote the five sources of innovation in the field of artificial intelligence: development policy, academic thought, theoretical foundation, top talents and industrial ecology. Based on the large-scale scientific research and computing practice environment of the Jiuding platform, KLCII has joined hands with many domestic AI chip companies such as Huawei, Kunlun Technology, Haiguang Information, Tianweixin, and Fangyuan Technology, as well as relevant scientific research teams such as Peking University, Tsinghua University, and the Institute of Computing of the Chinese Academy of Sciences, to jointly build an AI chip ecological laboratory, aiming to accelerate the ecological development and application of domestic AI chips.
In order to promote the wider application of large-model technology, the FlagOpen large-model technology open source system has reached ecological cooperation with many well-known enterprises such as China Mobile, Inspur Information, Kunlun Chip Technology, and Tiantian, and deployed FlagAI open-source tools on China Mobile Jiutian Bisheng Education, Inspur AI Station and other platforms.
Work together with partners to let Zhihu play a time difference in the field of large models. "Every morning at 10 o'clock, I, Zeng Guoyang, CTO of Face Wall, and some core engineers of Zhihu will have a morning meeting to discuss the progress of the big model and the depth of cooperation between us. In more than a month, the Chinese model of 'Zhihai Chart AI' cooperated by Zhihu and Face Wall was released. Li Dahai, CTO of Zhihu, recalled that now the Zhihai chart AI has been applied to the Zhihu hot list, which can capture, sort and aggregate the answers on the hot list, and show the answer outline to users.
A more down-to-earth example comes from the innovative cooperation between the Institute of Software of the Chinese Academy of Sciences and Calioma. "It is dynamic heating on demand, from full-space and full-time to time-sharing zone, temperature-sharing, directional quantitative and precise heating." Yang Siping introduced. Through the Kalioma gas radiant heating equipment, reduce heat energy consumption and reduce the heating cost of enterprises; Real-time collection of heating equipment data, monitoring equipment start and stop; The real-time database provides the basis for the analysis of heating energy consumption; Uniformly supervise heating equipment through the heat supply management platform, standardize and strengthen heat supply management, and improve energy-saving benefits.
AI + Everything
AI also makes takeaway couriers in Houshayu in Beijing's Shunyi district feel relaxed. "Now our site uses the latest generation of automatic delivery vehicles, more loading, longer mileage, during the working hours of the automatic delivery truck, we do not have to return to the station every time to pick up the goods, just wait for the delivery truck to be delivered at the gate of the community." It is understood that Meituan automatic delivery vehicles have been in regular operation in Shunyi District for more than three years, covering nearly 100 communities. At the same time, Meituan automatic delivery vehicles are also providing takeaway delivery services in many colleges and universities across the country. The latest generation has a loading capacity of 150 kg and a volume of almost 540 liters. At the same time, the cruising range has doubled, and the driving range of urban roads has reached 120 km.
The core product "Deep Pulse Score" displayed by Keya Medical at this Zhongguancun Forum is the landing practice of artificial intelligence + medical care.
"Deep Pulse Score" is the world's first product using deep learning technology for coronary physiological function evaluation, which adopts the company's self-developed DeepFFR technology, which can quickly, accurately and non-invasively analyze blood flow reserve fraction, and the analysis time only takes 10 minutes, and the FFR value of any point of coronary blood vessels can be observed. Further assessing whether coronary stenosis leads to myocardial ischaemia can help reduce the cost of invasive testing and avoid unnecessary interventions.
Yan Yuhua, industry director of Ronglian Cloud GTM, told a case of artificial intelligence + automobile. "Hyundai Motor hopes to improve its service model, 'customer-centric', build an application intelligent call center system, and the customized automotive industry customer service center solution provided by Ronglian Cloud can increase agent production capacity, improve agent service quality, integrate omni-channel customer contact trajectories, reduce invalid workload, and improve outbound call efficiency through the application of intelligent agent assistance system."
For the future development and application of artificial intelligence, Yin Qi, CEO of Megvii Technology, mentioned two directions: one is "AI in Digital" (AI in the digital world), that is, the wave of technology represented by AIGC (artificial intelligence generated content), which transforms the digital world by reconstructing content productivity. The second is "AI in Physical", that is, companies represented by Tesla make the physical world better by building different forms of robot carriers.
200 years ago, the work of carrying water from the underground largely disappeared, because the invention of the steam engine triggered the first industrial revolution, and the first application of the steam engine was how to use machines to bring mining water up. 100 years ago, the coachman's job disappeared, and in a 1913 photograph on Fifth Avenue in New York, the road was almost full of cars, only one carriage. Today's younger generation has not seen a typist, but the world's population and GDP per capita continue to grow at a high rate.
Every industrial revolution brings an increase in productivity, this time led by artificial intelligence.
Beijing Business Daily reporter Wei Wei