The deep winter is very cold, but 2021 in privacy computing is very hot.
This year, privacy computing is visible to the naked eye and has become the focus of the spotlight.
The capital market is booming, and hundreds of millions of financing are gathering together. With the integration and development of technologies and systems such as special chips, encryption algorithms, white boxes, and data trusts, global privacy computing has risen. At the same time, we can also glimpse the surging entrepreneurial undercurrent. More than 70 privacy computing companies were established this year. Blockchain companies, artificial intelligence companies, security companies, etc. are all laying out their plans one after another, aiming at the target of data. Large-scale applications have started, but the tug-of-war of competition is also starting. , is a test of survival or perishing.
This year, the construction of the data element market, data compliance and transactions kicked off the journey.
The "Data Security Law" and "Personal Information Protection Law" were officially implemented, and the Beijing International Big Data Exchange and Shanghai Data Exchange were established one after another... These many crucial nodes and events are forming a huge driving force, pulling With the industrial structure going deep.
The end of the year may be a good time to reflect and examine.
An annual long scroll about a series of keywords such as data, capital, privacy computing, and compliance is slowly unfolding, and the protagonists in the context of innovation will also appear one by one, connecting and blending into different cross-sections to form a computing power think tank 2021 in my eyes.
secondary title
2021 is recognized as the first year of data compliance
A series of regulations were issued and implemented this year:
01
On September 1, 2021, the "Data Security Law of the People's Republic of China" was officially implemented. This is a basic law in the field of data security in my country and an important law in the field of national security. Legal provisions have been made in terms of classification and grading, and important data exportation.
02
In October 2021, the Central Committee of the Communist Party of China and the State Council issued the "National Standardization Development Outline" and proposed that by 2025, the standard supply will be transformed from government-led to government and market, and the use of standards will be transformed from industry and trade to the entire economy and society. Strengthen standard research in key technical fields, and carry out standardization research in the fields of artificial intelligence and quantum information.
03
On November 1, 2021, the "Personal Information Protection Law of the People's Republic of China" officially came into effect. This is my country's first systematic and comprehensive law dedicated to the protection of personal information. Together, we will comprehensively build a legal framework in the field of information and data security in China.
04
Highlights of the Personal Protection Act:
(1) "Information-Consent Rules", which protect individuals' right to know and decision-making, and the right to restrict or refuse others' access to information during the entire life cycle of information collection, storage, use, processing, transmission, provision, disclosure, and deletion. Personal information is processed without the exception of obtaining the consent of the individual. For example, it is necessary to perform statutory duties and obligations, respond to public health emergencies, implement news and public opinion supervision for the public interest, etc.
(2) Gatekeeper Clause: Large-scale network platforms (with a large number of users and complex business types) should establish an independent agency to supervise the protection of personal information, establish a sound personal information protection compliance system, and accept social supervision.
(3) The right to portability of personal information: "number portability" of data.
(4) Regulate the behavior of state agencies: follow the principle of "clear purpose + minimum scope" in personal information processing.
(5) Maximum penalty: For serious violations of the law, the operator can be fined up to 5% of the previous year's turnover.
05
Local regulations: On March 29, the "Anhui Province Big Data Development Regulations" was released; on June 29, the "Shenzhen Special Economic Zone Data Regulations" were released; on July 30, the "Guangdong Province Digital Economy Promotion Regulations" were released; on August 2 "Zhejiang Provincial Public Data Regulations (Draft)" (draft for comments) was released; on September 30, the "Shandong Province Big Data Development Promotion Regulations" was released; on November 25, the "Shanghai Municipal Data Regulations" were passed.
There are also some golden sentence points of view, in deep insight:
01
Data classification and grading is the starting point of data protection and the premise of data security governance. As a key link in data security, it is not only the basis for the formulation of the internal management system of the organization, but also the support for the implementation of the technical tool system. It is proposed that according to the impact and importance of data on national security, public interests, or the legitimate rights and interests of individuals and organizations, data is divided into general data, important data, and core data, and different levels of data take different protection measures.
——Source: China Academy of Information and Communications Technology "Technical Requirements and Test Methods for Data Classification and Grading Tools"
02
Hierarchical governance of data is an important direction for the development of privacy computing technology. According to the requirements of business forms, hierarchical computing methods are implemented for data. It means that the calculation methods are different, and there is no classification with different degrees of privacy protection.
secondary title
Trading pilots, the business model has begun to take shape
Data transactions are at the right time, moving towards 2.0 upgrades, infrastructure and transaction rules are exploring a feasible path:
01
On March 31, 2021, Beijing International Big Data Exchange, the first data exchange in China based on the new transaction paradigm of "data available but not visible, use controllable and measurable", was established, providing a large-scale use of privacy computing for data transactions. The platform, based on the full-chain transaction service system supported by blockchain and privacy computing technology, will provide market participants with professional services such as data cleaning, supply and demand matching, legal consultation, value assessment, and ownership certification. As the basic transaction object, we will focus on solving the pain points of data transactions from five aspects: technology, model, rules, risk control, and ecology.
02
On November 25, 2021, the Shanghai Data Exchange was unveiled and established, focusing on key common problems such as data rights confirmation, pricing, mutual trust, admission, and supervision, and formed a series of innovative arrangements. The five "national first releases" of the digital business system, data transaction supporting system, all-digital data transaction system, data product registration certificate, and data product manual have realized full-time listing, global transactions, and full traceability. Through data product registration certificates and data transactions The issuance of vouchers realizes one count and one code, which can be registered, counted, and censused, which clarifies the direction for solving the "five difficulties" of data transactions.
03
On December 1, 2021, Shenzhen Data Exchange Co., Ltd. completed the industrial and commercial registration (referred to as Shenzhen Data Exchange). As an important entity that promotes the market-oriented allocation of data elements, Shenzhen Data Exchange will adhere to the functional positioning of public services, taking into account market-driven and compliance Development, to promote the aggregation of data resources and transaction flow.
04
Principles of data trading: Shanghai Data Exchange clearly regards "non-compliance, no listing, and no scenario, no trading" as the basic principle of data trading. Data trading compliance should at least cover: compliance of trading entities, compliance of data products, and trading process compliance.
05
Data productization should at least include: the legitimacy of the data source, the tradability and circulation of the data product itself. The current mainstream data trading platforms can be roughly divided into three trading modes: data package delivery mode, API delivery mode, and data hosting mode.
These views try to clear the air:
01
Today is the era of data capital, data financing will become an important model in the future, and the data trust system is a key option for building a credible data economy and achieving common prosperity. ——Zhong Hong, Director of Digital Economics Research Office, Technology Innovation Research Center, Tsinghua University
Source: Computing Power Think Tank "Shanghai Data Exchange was inaugurated today! What can a data trust do? "
02
Privacy issues are a major obstacle to data openness and data transactions. Only when privacy issues are resolved can ownership issues be resolved logically. ——Guo Bing, Deputy Director of Hangzhou Yangtze River Delta Big Data Research Institute
secondary title
The tide of digital business rises, and the deep water goes hard
01
On November 25, 2021, the Shanghai Data Exchange launched the digital business system, building a new "digital business" business model, covering data transaction subjects, data compliance consulting, quality assessment, asset evaluation, delivery and other fields. "Digital business" Covering various economic entities such as data discoverers, value enablers, connectors, and service providers, including suppliers and demanders of data products as data trading entities, and law firms responsible for compliance review of data products, There are also institutions responsible for asset assessment of data products, as well as entities that assist in the delivery of data products. Privacy computing vendors may play multiple roles, being both data product suppliers, demanders, and service providers.
——Source: KWM Research Institute "Imagination of Data Transactions and "Shanghai First Release""
02
"Digital quotient" refers to the economic entity that uses data as the main object of business activities or the main raw material for production. The release of the "three types of value" of data elements (discoverers, value empowerers, connectors and service providers) is the "digital quotient" 'duty of".
——Huang Lihua, executive deputy director of the National Engineering Laboratory of Big Data Circulation and Transaction Technology
03
Data transactions need to follow a complete procedural arrangement. From the scope covered by "digital quotient", it can be seen that at least a series of procedures such as quality assessment, asset assessment, compliance review, material submission, listing of data products, conclusion of transaction documents, and product delivery are required. process. Judging from the current public information, the compliance review conducted by law firms is a "must-have action". How should other procedures be arranged? Whether there are any regulatory requirements in the delivery process, etc., need to be clarified through the specification documents that will be formulated and published in the future.
secondary title
The scene is sinking, and the application is blooming
There is no fast, accurate and ruthless story about enshrining the gods on the road to the landing of privacy computing, and the magic of staking the land by Internet companies also does not work. Only when it sinks into the scene, can the application blossom and bear fruit.
01
In December 2019, the People's Bank of China launched a pilot project for financial technology innovation supervision. As of October 2021, 19 regions across the country have launched a total of 127 innovative supervision pilot projects, 13 of which involve privacy computing technology, and the application scenarios include financial consumers. Face information protection, product marketing, cross-border settlement, small and micro enterprise financing and credit risk control, etc.
02
In terms of financial business scenarios, privacy computing has further promoted financial business innovation, including clearer transaction model changes, transaction form innovations, and new compliance challenges.
——Source: Computing Power Think Tank "Unblocking Blockages, Data Sharing and Circulation Becomes a Powerful Medicine for Financial Innovation"
03
In the financial field, privacy computing is currently mainly used in two aspects of risk control and marketing. However, the impact of private computing on the financial field will not stop at these two aspects. After the combination of privacy computing and blockchain technology, more financial scenarios can be changed, such as cross-border payment, supply chain finance, etc.
——Source: "Opening a New Era: Report on the Application and Development of Privacy Computing in the Financial Field (2021)"
04
In the data silos and anti-money laundering business, take banks as an example. Previously, banks were very cautious in lending to SMEs due to their incomplete grasp of the operating data of SMEs. SME loans have become a major problem hindering market-oriented development. With the combination of privacy computing and blockchain and other technologies, banks can better collect relevant information and exchange parameters with peer institutions, and jointly calculate and model, providing a method to solve the problems of small samples and low data quality.
——Source: Computing Power Think Tank "Unblocking Blockages, Data Sharing and Circulation Becomes a Powerful Medicine for Financial Innovation"
05
Medical data security has always been the focus of supervision, including Internet medical data, wearable device data, and scientific research data. The requirements of system construction will inevitably bring about the interactive use of medical data, and privacy computing is an urgently needed application in the current compliance utilization of medical data. Technology. ——Feng Jie, Director of Information Department, East China Hospital
Source: Computing Power Think Tank "Medical: High-quality Human Data Sharing Scenario under the Epidemic | 2021 Privacy Computing Half Years"
06
If you only do a single application, you really don’t need privacy computing technology, but the trend we can see from many practical cases is that more and more different scenarios actually use the same data source, such as medical insurance anti-fraud, hierarchical diagnosis and treatment , medical record writing, syndrome prediction, etc. In the past, data was often reused and processed. Now the data base based on privacy computing can realize unified data opening, and then develop various applications based on such data base. In this system, the privacy computing platform provides the base and infrastructure for data elementization, enabling different cities and different departments to manage their own data, connect data, and collaborate together. The privacy computing platform allows data elements to Become usable and scalable at the technical level. ——Luo Zhen, CEO of Yifangjianshu
Source: Computing Power Think Tank "Using Technology to Unlock the Value of Data, Yi Fangjian's Dozens of Landing Cases Debuted! "
07
In recent years, the medical industry has become more and more accepting of privacy computing. The information departments of many tertiary hospitals basically do not need to do science popularization. The adaptation of privacy computing to medical applications is a future trend. It is not only necessary to strengthen the functions of the privacy computing base, but also It is necessary to find applications that cut into clinical problems and implement the applications based on the foundation. —— Zheng Hao, co-founder of Nuowei Technology, source: Computing Power Think Tank "Medical: Human High-quality Data Sharing Scenario under the Epidemic | 2021 Privacy Computing Half Years"
08
Privacy computing and the medical field are still in the "run-in period". The particularity of the medical industry puts forward higher requirements for the role of privacy computing. Compared with other fields, the methodology of the medical field is more complicated, and in many cases it has surpassed the scope of modeling , it is a challenge to support a complex methodology through the foundation of privacy computing, which involves a series of implementation and optimization points such as concurrency, accuracy, algorithm complexity, and privacy. —— Zheng Hao, co-founder of Nuowei Technology
Source: Computing Power Think Tank "Medical: High-quality Human Data Sharing Scenario under the Epidemic | 2021 Privacy Computing Half Years"
09
Compared with other industries, medical care has higher requirements for privacy computing. The test of the comprehensive ability of privacy computing is the balance of performance, security and precision, which can be achieved in three ways:
(1) The security level should be determined by specific scenarios to avoid over-protection or insufficient protection.
(2) On the premise of ensuring the safety protection level, determine the accuracy requirements in the data analysis process.
(3) Perform performance optimization in combination with privacy computing for the characteristics of medical data, characteristics of analysis methodology, multi-center characteristics, and high concurrency characteristics. —— Zheng Hao, co-founder of Nuowei Technology
secondary title
Privacy AI becomes a deterministic trend in the second half
After repeated privacy turmoil, today SenseTime is positioned as the first of the four AI tigers. However, the AI privacy issue is still unresolved like the sword of Damocles. It is foreseeable that artificial intelligence and privacy computing will become mutual needs in the future.
01
At the World Artificial Intelligence Conference in 2021, there is an obvious bright spot trend, that is, artificial intelligence is gradually showing a trend of integration and unity with other technologies including privacy computing, blockchain, Internet of Things, smart chips, etc. Among them, privacy computing is a computing theory and method for the protection of the entire life cycle of private information. Together with artificial intelligence, the core breakthrough is that the joint modeling and value sharing of data can still be realized under the premise of ensuring that "data is available and invisible". .
——Source: "How will AI + privacy computing set off a wave of business in the future?" ｜Computing Power Privacy Column》
02
Privacy computing is usually closely integrated with AI, and from a technical point of view, privacy computing is an important supplement to AI capabilities. AI is highly dependent on the data foundation. Large-scale and diverse high-quality data can train better models. Privacy computing provides data supplements for the continuous evolution of algorithms by solving the "link" problem of data.
——Xu Shizhen, Chief Architect of RealAI
03
At the product level, low replicability and poor versatility are major limitations facing the current privacy computing productization. The solutions are as follows: first, try to start from the requirements scenarios with mature specifications and less customization; second, embed privacy computing into existing mature products, such as privacy-protected databases and privacy-protected big data analysis engines. Under the latter idea, in the combination with AI technical capabilities, privacy computing can be regarded as AI middle platform 2.0, that is, adding privacy computing function modules to the original machine learning platform.
——Xu Shizhen, Chief Architect of RealAI
04
Efficient privacy AI can form a strong positive feedback network ecology. Once privacy AI enters self-evolution and continuously improves model efficiency, it will have strong error correction capabilities and compound interest generation capabilities. In the face of environmental changes, the safe flow of data assets allows personal privacy and business secrets to be traded under the condition of preventing leakage, while maximizing public interests.
secondary title
"Blockchain + Privacy Computing" becomes standard
In solving data problems, blockchain and privacy computing have the same goal by different routes, and they are also natural partners:
01
The combination of privacy computing technology and blockchain technology fundamentally balances the contradiction between data security and data value from the perspective of "physical dispersion and logical concentration" of data. This is where the value of privacy computing is most recognized. ——Zhang Yingchun, Vice President of Shudo Technology,
Source: Hashpower Think Tank "Half of the companies will increase privacy computing in 5 years, how to deal with sovereign technology blockade?" |Computing Power Privacy Column》
02
The right to control and benefit of the data has not really returned to the data producer himself. At the same time, because the cost of copying the data is close to zero, there is no proof of ownership, so the ownership and the right to use cannot be separated, and it is difficult to price. The data is confirmed through blockchain technology , Using privacy computing to protect data is an important technical means for the implementation of data production factors, so blockchain and privacy computing will definitely come to the fore. ——Meng Yan, vice president of Digital Asset Research Institute, founder of Token Thinking Lab
Source: Computing Power Think Tank "The First Domestic Privacy Computing Event Concludes, Blockchain and Privacy Collide, and the Value of Data Flow Can Be Expected in the Future"
03
Blockchain cannot solve privacy protection, and to solve this problem, we have to turn to cryptography. Only the solidification of privacy can confirm the rights, valuation, and pricing of data, so privacy computing is the business model that blockchain should have. Privacy computing represented by MPC will be the core infrastructure and methodology for pricing and rating data in the next generation. ——Sun Lilin, founder of Matrix Element
Source: Computing Power Think Tank "Dialogue with Wanxiang, Zhongan and other strongest brains, the golden age of privacy computing in 2020 may be strongly opened by the development of the alliance chain"
04
The blockchain is a ledger, and privacy computing is the final result. The combination of blockchain and privacy computing is mainly reflected in the market-oriented multi-party data sharing and co-governance mechanism. Mainly solve: the credibility of the participants, to ensure that the participants are authentic and not counterfeit; data asset governance issues, data assets need to be standardized before they can become commodities; participant data asset registration issues, this process is equivalent to commodity Put it on the shelf, so that potential partners can see the metadata of the data asset (that is, the necessary description information of the data asset) and how to use it; the credibility of the algorithm, the process of data processing must be visible to the data source, and the user should beware of The result contains private data; the authorization problem of calculation and the coordination problem of calculation process. ——Dr. Sun Lin, Senior Blockchain Expert, Data Intelligence Division, China Unicom
secondary title
The "big supermarket" of government affairs data has become an open carrier and has become a common practice
01
In 2021, the Sichuan Provincial Big Data Center is actively promoting the unified collection of data and overall governance work, creating a nationwide Sichuan standard node for government affairs data. In the future, it will gradually integrate the following data application platforms through the Sichuan Provincial Financing Big Data Service Platform and other scenario-based data application platforms. Various data security technologies represented by privacy computing provide government data compliance application services for various institutions, such as financial institutions. For example, the country's first large-scale application platform for privacy computing technology is being built, and offline incubation and cultivation services are being built to support it, so as to become a collaborative innovation application base for data elements.
——Source: Computing Power Think Tank "An article looking back at the evolution of privacy computing in the financial scene in 2021. What changes have taken place in the market? ｜2021 Privacy Computing Half Years》
02
On December 28, the Hainan Provincial Big Data Administration officially launched the "Data Product Supermarket" in Hainan Province. Based on the multi-party security technology and the technical architecture of federal computing, it realized member registration and certification, product listing and review, demand release, data development, Regulatory review, operation monitoring and other functions provide governments at all levels and various types of enterprises with data products that are good for governance, business development, convenience for the people, and business interests.
——Release of the pilot results of the development and utilization of public data resources in Hainan Province and the launch ceremony of the data product supermarket
03
For the first time, Zhejiang conducted a comprehensive inventory of the province's government information construction achievements over the past 20 years. 3,430 units at the provincial, city, and county levels participated. Based on a set of standards in the province, 10,129 government affairs digital application systems were analyzed and established The "integrated digital resource system" comprehensively surveyed digital resources such as information infrastructure, public data, application systems, and algorithm components of the province's government affairs system, forming an intelligent "general ledger" of digital resources. In this "big supermarket" "In ", all localities and departments can apply in the form of a "shopping cart", and digital resources can be efficiently shared, developed and utilized across departments, regions, and levels.
——Jiang Ruzhong, deputy director of Zhejiang Provincial Big Data Development Administration, source: 2021 World Internet Conference Wuzhen Summit
