Realize that data is available and invisible, taking into account security and application——
Privacy computing expects a benign ecology
Our reporter Chen Guojing
In the era of digital economy, with the rapid development of industries such as artificial intelligence and big data, the role and importance of data elements have become increasingly prominent.
At the same time, a large number of user data and information are mined and collected, and there are frequent user data and information leakage incidents in the financial field.
With the implementation of the "Data Protection Law" and "Personal Information Protection Law", how to protect data security, give full play to the value of data assets, and efficiently link multi-party data has become an urgent issue in the financial industry.
Break down data silos
Last year, the "Data Security Law" and "Personal Information Protection Law" were implemented successively, which put forward stricter requirements for data security and personal information protection.
At the same time, the financial management department has strengthened the law enforcement inspection of data security in the financial industry.
Under the trend of strict supervision, it is urgent for the financial industry to strengthen the security and compliance of data applications.
One side is security, the other side is application, how to balance the two?
The "Overall Plan for the Comprehensive Reform of the Market-Based Allocation of Factors" released on January 6 this year mentioned that it is necessary to establish and improve data circulation and transaction rules, and explore the transaction paradigm of "raw data does not go out of the domain, data is available and invisible".
Privacy computing may become a key technology.
Private computing is a technology and system in which two or more participants jointly compute and perform joint machine learning and analysis on their data through collaboration without revealing their respective data.
Under the privacy-preserving computing framework, the data of the participating parties is not available locally, and "data availability is invisible".
Yang Qiang, academician of the Canadian Academy of Engineering and the Royal Canadian Academy of Sciences, said that privacy computing can aggregate small data, ensure data security in the form of "data availability is invisible, data immovable value moves", give full play to the value of data, and further enhance the financial core Operational capacity.
"Under the tide of the digital economy, the increasingly strict regulatory environment has put forward higher requirements for enterprises and institutions to seek a reasonable balance between data protection and the rational use of data value." The international consulting agency IDC released "IDC Innovators, Privacy Protection Computing , 2022" report pointed out that driven by the dual needs of data fusion applications and customer privacy protection, as a key technology to realize the immovable value of data, the application of privacy-preserving computing can ensure that the data of the participants is not local, and the data security is protected. At the same time, the cross-domain cooperation of multi-source data is realized, which provides a feasible idea for solving the problems of data protection and integration application.
landing financial scene
At present, the financial industry is one of the most mature industry scenarios for the commercial application of privacy computing.
The current situation of high requirements and strict supervision of data security and privacy protection in the financial industry has greatly promoted the implementation of privacy computing in financial scenarios.
What exactly can private computing do for the financial industry?
According to industry experts, privacy computing can help financial institutions to further improve the marketing conversion rate and risk control efficiency.
For example, a bank calls 5,000 users every day to market credit products. In the past, the method was blind selection. The customers who received the call probably did not need the product, and the customer conversion efficiency was not high.
In the learning mode of private computing, data can be used as customer portraits, and customers who are more in line with product positioning can be selected from users, thereby greatly improving the conversion efficiency.
This effect has been verified in practice.
Especially in the context of more open financial business forms, privacy computing technology is becoming a "rigid need" in the financial industry.
At present, open banking is an important direction for the development of the banking industry, and data and data value sharing are the basic characteristics of open finance.
For example, in the banking risk control scenario, traditional banks conduct pre-loan analysis based on historical repayment information, credit data and third-party data.
However, this traditional method is not accurate, and there are multiple problems such as lack of data dimension and insufficient data.
A bank risk control model that adopts private computing can utilize third-party data to improve the effectiveness of the model and balance data security and privacy at the same time.
Another example is in the field of anti-money laundering, privacy computing can exchange encryption parameters and jointly model, effectively solve the problems of few anti-money laundering samples and low data quality, and form a robust and feature-rich intelligent model.
By invoking the jointly established model, the anti-money laundering capability of financial institutions is greatly improved.
The "Financial Technology Development Plan (2022-2025)" released by the People's Bank of China at the beginning of this year proposed eight key tasks for the next eight years, with particular emphasis on "data".
The plan proposes to build a green and high-availability data center, deploy an advanced and efficient computing power system, and further consolidate the "digital base" for financial innovation and development.
On the premise of ensuring security and privacy, promote the orderly sharing and comprehensive application of data, fully activate the potential of data elements, and effectively improve the quality and efficiency of financial services.
Connectivity is a consensus
At present, the main participants of domestic privacy computing include large Internet companies such as BAT and independent technology manufacturers of privacy computing such as Nebula Clustar, Huakong Qingjiao, and Shumu Technology.
In addition, some financial institutions are also exploring through self-research or cooperation with privacy computing vendors.
IDC China financial industry market analyst Wang Chen believes that at present, major manufacturers in the market are diverse in terms of routes and positioning, and different manufacturers compete and complement each other.
Next, in the construction of an orderly circulation system of data elements, more data participants are required to promote the improvement of norms and standards, break through the barriers to data circulation caused by technical differences, and ensure that technologies and applications meet the requirements of penetrating supervision , so as to form a benign ecosystem in which different subjects such as data providers, data demanders, technology providers, and regulators in the data chain participate and cooperate in an orderly manner.
At present, in order to break through the barriers caused by technical differences and avoid the formation of "computing islands", interconnection has become a consensus in the industry.
The China Academy of Information and Communications Technology, together with a number of institutions, has compiled the overall framework of the series of standards for "Privacy Computing Cross-Platform Interconnection", which involves communication specifications, interconnection protocols, and application specifications.
Open source is also considered as one of the realization paths of privacy computing interconnection.
Open source is based on the logic of "open source, open technology, inclusive technology", which helps software development to reduce costs and increase efficiency, accelerate iteration, promote industrial collaborative innovation, mine data value, and provide impetus for the development of various industries.
The "Privacy Computing White Paper (2021)" released by the China Academy of Information and Communications Technology pointed out that open source collaboration is accelerating the technological iteration of privacy computing.
As a key basis for ensuring cross-agency data security cooperation, privacy computing is destined to include an open source model.
The cost advantage of open source is not only reflected in technology reuse, lowering the development threshold, but also in the agility of problem discovery and repair, which is conducive to accelerating technology upgrades and commercialization of projects.
Data shows that 55% of domestic privacy computing products are based on or refer to open source projects.
Relevant analysis believes that the key to the development and large-scale application of privacy computing technology is to build an ecosystem that allows all parties in the ecosystem to connect and collaborate.
Open source projects naturally possess technical openness and iterative capabilities, and the interconnection of all parties has become the key to ecological construction.
Compared with the current "a hundred flowers blooming" among different privacy computing vendors, the interconnection based on the same open source framework will be more conducive to the formation of the privacy computing industry ecology.