In the era of digital economy, data is increasingly showing its important value. When data becomes as important a factor of production as land, labor, capital and technology, two problems follow: one is how to protect the data rights and interests of all social parties when enterprises or institutions provide services through big data mining; Second, the amount of data owned by a single enterprise, research institution or government department is limited after all. In some cases, it is difficult for everyone to share data frankly, objectively creating "data islands". Can there be a mechanism to share data with each other without revealing core secrets? As a technology to balance data utilization and privacy protection, privacy computing is emerging quietly and playing an increasingly important role. Big data "kills", privacy computing says "no" At present, all walks of life are promoting digital transformation. Digitalization is not only the only way for enterprise development, but also the inevitable trend of economic and social development. With the Internet, big data, cloud computing, artificial intelligence, blockchain and other technologies accelerating innovation and increasingly integrating into the whole process of economic and social development, the value of data has become increasingly prominent. A well-known case in the e-commerce field so far is that families who buy milk powder will probably buy sweeping robots next. This positive correlation mined by big data indicates that buyers may just have a baby at home. Financial institutions have also started to use big data to draw user portraits for a long time. In addition to the basic information such as unit, income and education background, the data even covers tens of thousands of dimensional combinations such as access occasions, reading types and web browsing frequency. Similar "big data portraits" can certainly improve the marketing efficiency of e-commerce, enterprises and financial institutions, and reduce customer acquisition costs and operational risks, but where is the privacy boundary of data mining has also triggered many discussions. According to the 49th statistical report on China's Internet Development released by China Internet Network Information Center (CNNIC) in February this year, by December 2021, the number of Internet users in China had reached 1.032 billion, and the Internet penetration rate had reached 73.0%; The average Internet user in China spent 28.5 hours online every week, an increase of 2.3 hours over December 2020. Obviously, the Internet has been deeply integrated into people's daily life, and data has changed people's production, life and consumption patterns, which puts forward higher requirements for the data control ability of all social parties. In this context, information security and privacy protection have been raised to an unprecedented height in the process of promoting digital transformation. The personal information protection law of the people's Republic of China, which came into force on November 1, 2021, is the first special law on personal information protection in China. Together with the data security law of the people's Republic of China and the network security law of the people's Republic of China, which have come into effect before, they form a "troika" in the field of data protection. How to balance information security and privacy protection in the process of digitalization? Private computing that makes data "available but invisible" may provide a solution. Yang Guang, research director of Shanghai tree blockchain Research Institute, explained that, in short, privacy computing is like a black box. People do not need to see how the black box operates to get the calculation results, "which is equivalent to people sharing the data value and protecting their digital rights and interests." At the technical level, privacy computing is almost applicable to all links of the current Internet, so that data can be used and privacy can be protected. Therefore, its emergence is regarded by the industry as a "good medicine" to "cure" a series of problems such as excessive personal information collection, big data "killing" and privacy disclosure. Want big data, but don't want to be seen The larger the amount of data and the higher the quality, the more accurate the modeling training and prediction can be. However, not everyone is willing to disclose their own data or models. Firstly, the data is too easy to be copied, and they are afraid that taking out the core data will reduce their competitiveness. Secondly, the data in some fields are not suitable for disclosure. For example, cloud computing is a way to save costs and improve computing power, but one of the reasons why some enterprises do not want to "go to the cloud" is that "going to the cloud" means giving their own data and algorithm models to cloud service providers. If there is a technology that can restrict the permission of cloud service providers to view computing data, the enthusiasm of enterprises to "go to the cloud" will be greatly improved. Similarly, among upstream and downstream enterprises or competitors, their data are often highly correlated and complementary. If they can get together, they will play a stronger role than fighting alone. However, under the condition of ensuring its own data security, no enterprise is willing to take out its own core data. Yuanchen, associate professor of the school of Cyberspace Security of Shanghai Jiao Tong University and chief cryptology scientist of the tree of light technology, said that the key to cracking the "data island" is to realize data sharing while protecting the interests of all parties. Privacy computing is one of the recognized directions in the industry and has developed rapidly in recent years. Generally speaking, privacy computing is designed to meet the needs of "big data without being seen". For example, when sending express, the "send before pay" mode can save at least 5 seconds and at most 1 minute per order. In the context of the epidemic, with the use of privacy computing technology, the tree of light technology has tripled the target customers of SF express "first send then pay". How does this work? Yuanchen explained that the target customers of "post first and pay later" are those users with good financial credit. However, SF did not have such user information, but the bank already has relatively perfect customer credit information. However, in general, the bank will never give the core information to others. As a leading privacy computing company in China, the tree of light has previously provided a privacy computing solution for China UnionPay, so that they can only tell SF whether the user can become a target customer of "post before pay" without disclosing the user's information. Thus, the same data has played a greater role. In the field of industrial digital intelligence and precision marketing, privacy computing also has great potential. For example, an enterprise hopes to find potential customers similar to existing customers and expand customer groups by means of database comparison. However, enterprises do not want to open their core user list to third parties. "If you don't give it to me, how can I find it?" Privacy computing can help enterprises solve this dilemma. The prosperity of Internet applications leads to a new track Every factor of production must have a matching sharing and trading mechanism. The booming privacy computing undoubtedly provides a technical support for the underlying logic of the digital economy. Yuanchen introduced that privacy computing is divided into hardware scheme and software scheme. The hard scheme is the trusted computing tee (trusted execution environment), and the software scheme is divided into secure multi-party computing and federated learning privacy computing. At present, China is in the first echelon in this field. "Under the same conditions, the combination of software and hardware is better than software alone." In recent years, Yuanchen has maintained an average frequency of 1-2 papers collected every year at international top cryptography conferences. He is also the scholar who has published the largest number of papers on privacy computing at the top three international cryptography conferences. In terms of hardware, the secret computing server, which supports privacy computing, came into being. It refers to a server that can not see the operation data in the operation process and only output the operation results under the chip transformation and hardware support. It is suitable for large-scale data operation. As a newly emerging segment track, privacy computing has developed rapidly in China in recent years. Globally, not only the Internet giants have launched relevant departments, but also the number of start-ups is very considerable. They often exist as trusted third parties of various units. For example, owkin, a start-up that provides a privacy computing platform for the world's top pharmaceutical companies, actually acts as an intermediate node linking hospitals, scientific research institutions and pharmaceutical companies. "Privacy computing is an application-oriented technology that needs to be supported by a prosperous and powerful digital industry. In this regard, China has a strong digital payment, has fully developed Internet applications, and laws and regulations have been issued to protect digital rights and interests. It should be in the forefront of privacy computing." According to Yuan Chen, the vigorous rise of privacy computing in China is a vivid portrayal of the development of China's digital economy. "Shanghai has unique conditions to develop private computing." Zhangjiachen, founder of the tree of light technology, said that this is also the reason why the company is headquartered in Zhangjiang. At the end of last year, the Shanghai data exchange was inaugurated, and the tree of light technology became the first technical service provider of the exchange. In her opinion, Shanghai has gathered a large number of semiconductor enterprises and Internet industries, as well as a large number of financial institutions and biomedical enterprises. Under the "escort" of privacy computing, more new business forms and models will be created to make the digital economy run faster and more stable. (Xinhua News Agency)