Since the “big data” first entered the government work report in 2014, the Chinese market has started a big data “sweeping” boom. The China International Big Data Industry Expo (abbreviated as: Data Expo), which started in 2015, to the “Never Ending Data Expo” during the COVID-19 pandemic in 2020, has been baptized by the “big data expo” which lasted for 5 years and 6 times. The concept of "data" is extended to the streets. In the past five years, China's big data technology industry has also flourished and formed a strong technical support system. According to the China Academy of Information and Communications Technology, as of October 2020, there are more than 3,000 active big data companies in my country.
After 5 years of development, "big data" has now gone from a pure technical architecture and technical system to social infrastructure. In 2020, "New Infrastructure" will define "Big Data Center" as an important construction content of new digital infrastructure. At the end of 2020, Shanghai issued the "Opinions on Comprehensively Promoting the Digital Transformation of Shanghai's City", clearly proposing to "rebuild the social operation process of the digital age", especially to guide enterprises to realize a data-based "decision-making revolution" and guide the market to reshape digital The cognitive ability and thinking mode of the times promote the government to reengineer data-driven processes.
However, according to the "Big Data White Paper (2020)" of the China Academy of Information and Communications Technology: Only 56% of the data in enterprise operations can be captured in a timely manner, and only 57% of the data is used, and 43% of the collected data is not. Activation means that only 32% of the value of enterprise data can be activated. Although in the next two years, enterprise data will maintain rapid growth at a rate of 42.2%, how to activate the value of data and truly "rush for gold" from big data has become the top priority of big data in 2021.
Data Fusion and Data Value Mining
Data fusion is of great significance to data value mining. Wu Hequan, academician of the Chinese Academy of Engineering, said at the "Never Ending Digital Expo" 2020 series of activities-"Big Data Industry Ecological Innovation and Development Summit", data integration and utilization require standards and regulations to achieve data visibility, easy understanding, and Data linkability, data credibility, data interoperability, data security. At the same time, the fusion and utilization of data faces the challenge of establishing a mathematical model, and big data mining faces the challenges of computing power and algorithms, data sample accuracy, small data, and human and data fusion.
Wu Hequan believes that there are many scenarios that require data fusion applications in smart city management and the industrial Internet. Multiple heterogeneous data fusion will revitalize data, develop data value through data mining, and play the role of data as a production factor. Data mining and AI analysis need to face the challenges of massive processing capabilities, cloud-side collaboration, modeling, small data, human and data fusion, data security, privacy and business secret protection, etc., and need to be based on many aspects of basic theory and engineering practice. Research on the value mining of data elements and develop more big data and AI analysis technologies.
In cross-enterprise data fusion, it is necessary to ensure that data can be shared but sensitive data is not leaked. A new data analysis mechanism needs to be established, and a virtual black box is used to keep data from being managed by the unit and authorized by other units. Call, the existing trusted computing environment method based on black box, but still need to prove the reliability of the third party.
In terms of innovative data circulation techniques, federated learning is an important machine learning framework. Federated learning is a machine learning framework proposed for "data islands" and privacy and security issues in data sharing. The traditional way is to move hard copies of data to a trusted third party, while federated learning hopes to achieve the freedom of various enterprises and institutions. If there is data that does not go out locally, a virtual shared model is established through parameter exchange under the encryption mechanism. The data itself does not move nor leak privacy or affect data compliance.
Data-agile economy
Countries in the world are strengthening the development of big data, launching national data strategies, and exploring the road to data development. Among them, the EU's data strategy is extremely specific and representative. As the EU is an economic union composed of more than 20 countries, the various policies formulated by the EU are more forward-looking for other single countries or economies, taking into account the coordination, balance and balance between different markets, countries and regions. Inclusiveness is also more representative.
According to the "Big Data White Paper (2020)" of the China Academy of Information and Communications Technology, the EU is committed to balancing data flow and widespread use in response to future development. It hopes to establish a single data market to ensure that Europe will occupy a leading position in the future data economy. In February 2020, the European Commission announced the "EU Data Strategy", which set a clear vision for the development of EU data-Europe will become the most attractive, safest and most dynamic data-agile economy in the world by 2030 body. That is, on the premise of maintaining a high degree of privacy, security and ethical standards, fully explore the value of data utilization to benefit the economy and society, and ensure that everyone can benefit from the data dividend.
In May 2020, Helsinki EU Office members jointly released version 2.0 of the White Paper "Data Agile Economy: From Passive to Active, Better Serving the Society", suggesting the transition from passive use of data to active use The legislative framework for data transformation: innovative cross-departmental use of data to play a major role in more efficient, proactive and personalized public services; close cooperation between different ecological partners, including public institutions, RDI organizations, private companies and NGO non-profit organizations play a key role in unleashing the potential of digital transformation; let everyone use and use personal data in their own way, so as to bring benefits to individuals and contribute to society; through better data management and updates High-level data culture and digital skills to achieve data trust and empowerment of the people.
Knowledge Graph and Decision Intelligence
With the development of big data, enterprises and public institutions increasingly need to effectively link different data to form new dynamic knowledge to assist enterprises and public institutions in their decision-making. This requires the use of graph databases, graph computing engines, and knowledge graphs. Knowledge graphs are important application scenarios for graph databases and graph computing engines. According to the ranking analysis of DB-Engines, the interest of graph database has increased by 10 times from 2013 to 2020, and the growth of interest ranks first, far higher than other databases or data engines. User portraits and credit files are new application scenarios of knowledge graphs.
Manbang Group is a big data unicorn company that emerged in Guizhou. Manbang is an intelligent capacity platform in the field of highway logistics. It connects truck drivers and cargo owners to dual-end users. Through the intelligent recommendation of big data, it helps both parties quickly complete transportation transactions, improve transportation efficiency, reduce transaction costs, and reduce truck empty driving rates. At the end of November 2020, following the $1.9 billion financing in 2018, Manbang announced the completion of a new round of financing of approximately $1.7 billion. As of November 2020, Manbang platform has certified more than 10 million drivers and more than 5 million certified shippers.
The logistics industry is an industry with a weak credit system. Previously, a dispute occurred for every four transactions completed, and the cost of disputes was very high. Manbang established a Van Gogh system, established a credit profile portrait for every user, driver, and cargo owner, including more than 200 dimensions of indicators, and established a blacklist of untrustworthiness, with the National Development and Reform Commission and the National Information Center "Credit China" The system is connected and joint disciplinary action is implemented, and it is also included in the central bank's credit investigation system to create a "second ID card" for truck drivers, reducing the industry dispute rate from 25% to 3%.
At present, domestic large-scale cloud vendors such as Alibaba Cloud, Huawei, Tencent, Baidu, and some start-up companies are laying out graph databases, graph computing engines, and knowledge graphs. In particular, knowledge graphs have begun to be applied to finance, industry, energy, etc. Industries and fields. Knowledge graph is becoming an important technology platform and tool for corporate decision-making.
Industrial Internet of Things speeds up
The Internet of Things is an important source of big data. The traditional concept believes that the consumer Internet of Things is the main source of big data on the Internet of Things, but with the rapid development of the Industrial Internet of Things, the Industrial Internet of Things is surpassing the Internet of Consumer Things and becomes the main source of big data on the Internet of Things. According to the "Internet of Things White Paper (2020)" of the China Academy of Information and Communications Technology, as the Internet of Things accelerates its penetration into various industries, the industry’s informatization and networking levels continue to improve, and the proportion of industrial Internet of Things connections will increase.
According to GSMA Intelligence, the number of connected industrial IoT devices will exceed the number of consumer IoT devices in 2024. In China, the industrial Internet of Things and the consumer market account for half of the number of Internet of Things connections in China in 2019. It is expected that by 2025, most of the growth in the number of Internet of Things in China will come from the industrial market, and the number of industrial Internet of Things connections will account for the overall Of 61.2%. Smart industry, smart transportation, smart health, smart energy and other fields will most likely become the fastest growing areas of industrial IoT connections.
The big data processing of the Industrial Internet of Things involves edge computing. IDC, a market research organization, predicts that in the future, more than 50% of data will need to be stored, analyzed, and calculated on the edge side. By 2024, the global edge computing market will reach 250.6 billion US dollars. In the Chinese market, 5G, AI, smart transportation, new energy vehicle charging piles, and industrial Internet in the 2020 new infrastructure are all technologies or scenarios related to edge computing.
According to the "Edge Computing and Cloud Computing Collaborative White Paper 2.0" jointly released by the Edge Computing Industry Alliance (ECC) and the Industrial Internet Industry Alliance (AII) in December 2020, the use of AI technology for edge data processing is facing challenges: AR, VR, Interactive live broadcasts, video surveillance, and other scenarios dominate unstructured data. The challenges are large data volume, large resource usage, high real-time requirements, and difficulty in data labeling. In industrial scenarios, IoT structured data is dominant, and the challenge lies in fewer samples and coldness. Start-up and request models are interpretable and reliable. The key technologies of the edge-cloud collaborative AI framework include: incremental learning, federated learning, and joint reasoning.
Data security continues to rise
Data sharing and circulation is an important direction to release the value of data. Shen Changxiang, an academician of the Chinese Academy of Engineering, delivered a keynote speech on "Using Active Immune and Trusted Computing to Ensure High-quality and Healthy Development of the Digital Economy" at the "Never Ending Digital Expo" 2020 series of activities-"Big Data Industry Ecological Innovation and Development Summit". He said that the development of data "science" has gone through the processes of numerical calculation, data engineering, and digital society. Cyberspace has become the fifth largest sovereign domain after land, sea, air, and space. Without cybersecurity, there would be no national security.
Big data and the digital economy must adopt corresponding legal systems and related measures to ensure healthy development. The first is to change the calculation method to protect while calculating; the second is to build an immune system and change the security architecture; the third is to build a "triple" protection framework of "security office", "guard room" and "safe express delivery" for network system security; fourth is Dynamic credibility measurement, identification and control of the four elements of human operation access strategy (subject, object, operation, environment); the fifth is "risk analysis, accurate rating", "review and filing, standard construction", "perceived early warning, emergency response" Countermeasures, “strict evaluation, rectification and improvement” and other links are controlled throughout the entire process, with equal emphasis on technical management; six is that important information cannot be obtained by unauthorized persons, systems and information cannot be changed, attack behavior cannot be ignored, attackers cannot enter, The "six no" protection effects such as stealing confidential information, unintelligible, failure of system work, etc.
The "Big Data White Paper (2020)" of China Academy of Information and Communications Technology also pointed out that data circulation technology based on privacy computing has become the main idea for realizing data joint computing. In an environment where the demand for data compliance and circulation is strong, privacy computing technology is developing rapidly. Privacy computing realizes data fusion under the premise of protecting the data itself from external leakage, and brings the possibility of safe and compliant data circulation.
In summary: Entering 2021, big data has evolved from a purely technical system to a direction that integrates with the real economy and truly excavates and exerts the value of data. In particular, the new crown epidemic and new infrastructure have accelerated the rapid integration of big data and physical social infrastructure, and the rapid development of 5G and the Internet of Things has further increased the deep integration of big data and the real economy. As the Digital Expo is about to enter its sixth year, big data will truly penetrate into all aspects of the social economy and promote the next round of economic long-term cycle. (Text/Ningchuan)