Reporter Yang Junfeng.
On August 24, 2023, two technicians check the operation status of a lathe at the machining digital centralized control center of Baosteel Roll Technology Co., Ltd. in Zhonglou Economic Development Zone, Changzhou City, Jiangsu Province. Photo by Lu Shiqing (People**).
On October 21, 2023, in Wuhu City, Anhui Province, citizens experience virtual reality technology at the 11th China (Wuhu) Science Products Expo. Photo by Xinhua News Agency reporter Du Yu.
On November 23, 2023, in Hangzhou, Zhejiang Province, visitors learn about the application of Huawei's Pangu model at the 2nd Global Digital Expo 2023. Photo by Long Wei (People**).
In this day and age, data has become an important element of innovation. Artificial intelligence models, new material creation, biological breeding, and basic scientific research are all inseparable from the support of data.
The National Data Bureau and 17 other departments recently issued the "Data Elements" Three-Year Action Plan (2024-2026)", which clearly carries out the action of "Data Elements and Scientific and Technological Innovation", and expounds the focus of the combination of data elements and scientific and technological innovation from the aspects of promoting the orderly open sharing of scientific data, strengthening the construction and scenario application of high-quality scientific data resources, using scientific data to help cutting-edge research, supporting technological innovation with scientific data, supporting the development of large models with scientific data, and exploring new paradigms for scientific research.
From supporting basic research, to helping the development of cutting-edge technologies such as artificial intelligence, to promoting the transformation of scientific research methods, China's scientific and technological innovation is "multipliing" with the "east wind" of the three-year action plan of "data elements".
Build the "hardware and software".
It is one of the important goals of the "Data Elements Science and Technology Innovation" action to promote the orderly open sharing of scientific data, promote the interconnection and interconnection of all kinds of scientific data generated by major scientific and technological infrastructure and major scientific and technological projects, support and cultivate the construction of scientific databases with international influence, and rely on platforms such as the National Science Data Center to strengthen the construction of high-quality scientific data resources and scenario applications.
Build relevant "software and hardware", and all localities are actively deploying.
In 2024, Beijing will promote the implementation of a number of major projects such as computing power centers, data training bases, and national blockchain hub nodes.
Jiangsu will systematically promote the large-scale deployment of 5G and gigabit optical networks, support the construction of national Internet backbone direct connection points in Suzhou, and accelerate the layout of computing facilities such as intelligent computing power and edge computing.
Sichuan proposes to build digital information infrastructure in advance, accelerate the construction of national hub nodes of the "Eastern Data and Western Computing" project, build a provincial computing power dispatching service platform, and build an integrated computing network development system of computing power, storage capacity and transportation capacity.
Shandong proposes to deploy a high-performance intelligent computing center, coordinate the layout of general and vertical large-scale model computing power, and build more than 25 5A-level provincial-level new data centers, with the proportion of intelligent computing power reaching 30%, and build a "Shandong computing network". Support Jining to build Lunan Computing Center. We have carried out the "dual gigabit" network system project, created more than 500 typical application projects, and opened 40,000 new 5G base stations.
The "hardware" facilities are strengthened, and the "software" facilities also need to be upgraded.
The Internet is a platform for data circulation and aggregation, and the key to infrastructure in the era of digital economy. Mei Hong, an academician of the Chinese Academy of Sciences, said that it is necessary to accelerate the construction of new infrastructure such as data networking and data space.
In 2021, the Chinese Academy of Sciences released an open and general-purpose scientific data storage and publishing platform with international service capabilities, the Science Data Bank (ScienceDB).
Independently developed by the Computer Network Information Center of the Chinese Academy of Sciences, the Scientific Data Bank is a first-class associated data storage platform, which can provide efficient solutions for the aggregation, management, openness and sharing of relevant data, and provide a platform and service guarantee for the implementation of scientific research integrity, the cultivation of a sharing culture, the acceleration of data flow and the promotion of international cooperation.
Researchers can store and publish the scientific data they have collected and sorted out in the scientific data bank, which collects the scattered data resources of individuals and collectives by absorbing "data deposits" and "turning small money into big money, and dead money into living money", making it easier to discover, access, interoperate and reuse. At the same time, researchers can also upload the data to the scientific data bank before academic submission.
As of February 2 this year, the Scientific Data Bank has collected more than 8.2 million open datasets, with more than 700 million visits to the platform.
Develop large models.
The development of artificial intelligence models is another important goal of the "Data Elements Science and Technology Innovation" initiative.
Data elements The "Three-Year Action Plan (2024-2026)" proposes to support the development of large models with scientific data, dig deep into various scientific data and scientific and technological literature, build a scientific knowledge resource base through fine-grained knowledge extraction and multi-knowledge integration, build high-quality corpora and basic scientific data sets, and support the development and training of artificial intelligence large models.
In recent years, China has a good computing power foundation and a broad market in the field of large models, and domestic large models have frequently appeared and accelerated iteration. According to data from CCID Research Institute of the Ministry of Industry and Information Technology, at present, there are more than 19 large language model R&D manufacturers in China, of which 15 manufacturers' model products have passed the record.
With general capabilities such as language comprehension, logical reasoning, knowledge question answering, and text generation, these large language model products have been welcomed by users as soon as they were launched.
New breakthroughs have been made in scientific and technological innovation. The iFLYTEK Xinghuo cognitive model is at the leading level in the country. This is a sentence written into this year's Anhui Provincial Work Report.
iFLYTEK Xinghuo is a new generation of cognitive model officially released by iFLYTEK in May 2023. At present, iFLYTEK Xinghuo has been affirmed in many evaluations by the National Research Institute of Economic Research of the Development Research Center, the China Enterprise Development Research Center of the Xinhua News Agency Research Institute and other institutions, and is known as China's high-quality domestic model.
Only by building the large model on a completely autonomous and controllable platform can we firmly grasp the development initiative in the era of general artificial intelligence in our own hands. Liu Cong, dean of the iFLYTEK Research Institute, told this reporter. In October 2023, at the iFLYTEK Global 1024 Developer Festival, iFLYTEK announced that it would work with Huawei to build the "Feixing No. 1" platform, the computing power base of the domestic large model.
At present, based on the "Flying Star One", the iFLYTEK Xinghuo V35 completed training and was released on January 30. The upgraded iFLYTEK Spark v35. The ability in logical reasoning, language comprehension, text generation, mathematical answering, multimodality, etc. has been significantly improved. At the same time, iFLYTEK also released the Xinghuo voice model and the open source model.
Large models bring new opportunities for the development of voice technology. Liu Cong said. Enabling machines to learn, reason, and make decisions is the main job of cognitive models.
We believe that there may be the following four trends in the development of AI models in the future. Liu Cong told this reporter, "The first is multimodality and multilingualism. From the perspective of the future development of general artificial intelligence, the cognitive intelligence model is the core foundation, based on which other data such as voice, image, and ** can be aligned into a unified semantic space, combined with plug-in tools to achieve multimodal system presentation. The second is credible and explainable. This requires ensuring the source quality of massive data, the continuous optimization and iteration of the capabilities of the large model itself and the system solution, as well as the regulatory policies and laws and regulations issued by the state. The third is to develop in the direction of systemic innovation. With the support of large model capabilities, we need to combine multiple advantageous technologies to carry out systematic innovation and pay attention to the moat effect brought by it. The fourth is the integration of software and hardware and the development of national production. At present, iFLYTEK has invested in and deeply participated in the ecological construction of domestic AI chip software, and has made certain gains and progress on the training side and inference side. ”
Promote the integration of digital intelligence.
Intelligent search, keyword filtering, and access to the latest medical news ......With the convenience of big data and artificial intelligence, users can easily experience these services at their fingertips. In October 2023, the standard cloud sharing station of the big health industry chain in Taizhou City, Jiangsu Province was officially launched, providing enterprises with genuine current, real-time update, and more user-friendly information service support with millions of standard data.
This is Taizhou's innovative practice to promote the deep integration of digital and intelligent technologies and standards. Looking at Taizhou's pharmaceutical and health industry, from online to offline, from "laboratory" to "workshop", the achievements of "intelligent transformation and digital transformation" have moved to "production lines".
Walking into the small-volume injection production workshop of Jiangsu Dameng Pharmaceutical in Taizhou Pharmaceutical High-tech Zone (Gaogang District), the automatic production line operates in an orderly manner, and the drug production is efficient, accurate and stable; The production plants of Yangtze River Pharmaceutical Group have made the whole process of production more intelligent and digital through a series of "intelligent transformation and digital transformation" measures such as "machine substitution" and integrated management of information systems. Jiangsu Longfengtang Traditional Chinese Medicine has formed a complete set of modern solutions from the pre-treatment to the extraction of Chinese herbal medicines, and has created a model of "intelligent transformation and digital transformation" in the field of intelligent manufacturing standardization construction of traditional Chinese medicine process.
Taizhou's practice shows that at present, data has become an important element of the pharmaceutical and health industry, and digital technology has also become a necessary tool for the innovation and development of biomedicine. The integration of digital intelligence is indispensable for empowering the innovation and development of the pharmaceutical and health industry.
In the action of "Data Elements and Scientific and Technological Innovation", promoting the integration of digital intelligence is also one of the important measures. The "Data Elements" Three-Year Action Plan (2024-2026) proposes to support technological innovation with scientific data, focus on biological breeding, new material creation, drug research and development and other fields, and accelerate technological innovation and industrial upgrading with the integration of digital intelligence.
In recent years, the integration, iteration and diffusion of a new generation of digital intelligence technologies such as artificial intelligence, blockchain, deep learning, and the Internet of Things have penetrated into all aspects of R&D and design, manufacturing, and customer service, bringing full-angle, all-round, and full-chain transformation to production technology and production methods, comprehensively improving the level of automation, digitalization, and intelligence in the industry, and providing a key driving force for promoting the integration and application of digital intelligence and accelerating the formation of new quality productivity.
Qian Xiaojing, a professor at the School of Economics and Management of Northwest University, believes that data elements, as a new type of key production factor stored and processed by computer equipment in an immaterial form in the era of digital economy, have technical and economic characteristics such as non-competitiveness, low replication cost, non-exclusivity, and strong externalities, and can be reused by different subjects, and generate the integration value of "data + algorithm + computing power" through deconstruction and reorganization, convergence and integration, etc., which provides basic and important resources for consolidating the internal form of digital intelligence integration and accelerating the formation of new quality productivity.
On the one hand, the organic integration and application of data elements and traditional production factors enriches the expression form of digital intelligence integration, and shifts the mode of digital intelligence integration from geographical space to digital space. "On the other hand, data elements play a 'medium' role in the reconfiguration and recombination of traditional production factors, resulting in a new factor processing model and giving birth to a new form of digital intelligence integration." ”